JP2006189339A - Thin film shape measuring method and thin film shape measuring instrument - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a thin film shape measuring method and a thin film shape measuring instrument for minutely measuring the shape of each reflective surface of a physical object and its film thickness distribution on a quadratic plane. <P>SOLUTION: A personal computer 62 estimates a modulation amplitude Z<SB>bi</SB>and a phase α<SB>i</SB>with respect to each reflective surface of the physical object 50 included in a theoretical expression by minimizing an error function H where the error function H is the sum of squares of a difference between the value of a processed signal obtained from an electrically converted interference signal S(t) and the value of a processed signal that is the theoretical expression. From estimated values on the modulation amplitude Z<SB>bi</SB>and the phase α<SB>i</SB>, the position of each reflective surface of the object 50 is found, making it possible to minutely measure the shape of each reflective surface of the object 50 on a quadratic plane while excluding mechanical error factors as seen in the past. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a thin film shape measuring apparatus for simultaneously measuring the shape of each reflecting surface of a thin film such as a thin film on a semiconductor wafer, a plastic film, or a coating film, and the film thickness distribution on a secondary plane.

  In general, as a method for measuring the film thickness of an object with a thin film as an object, measurement methods using an incident angle scanning measurement method and a wavelength scanning measurement method are known. In the incident angle scanning measurement method, a monochromatic light is emitted from a light source such as a semiconductor laser toward an object while changing the incident angle, and the intensity change of reflected light from the object is measured, thereby measuring the object. The film thickness of the object is calculated. However, since the measurement is performed by applying monochromatic light to one point of the object, the number of measurement points is limited to one point.

  The wavelength scanning measurement method uses a multi-channel spectrophotometer (MCPD) to measure the amount of light reflected from an object while changing the wavelength of light emitted from the light source to the object using a halogen lamp or the like as a light source. The film thickness of the object is obtained. In this method, the measurement points are linear, but in order to measure the film thickness on a two-dimensional surface, it is necessary to scan the linear scanning points by some means. Moreover, although it can measure about a thin film of a single layer, there is a dissatisfaction that cannot be applied to a multilayer film having a plurality of reflecting surfaces. Furthermore, in the above two methods, the film thickness distribution of the object can be obtained, but the shape of the front and back surfaces of the thin film cannot be obtained.

Apart from this, a measuring apparatus using a white interferometer that can simultaneously measure not only the film thickness distribution but also the shape of the front and back surfaces of the film thickness is known. The principle is also shown in Patent Document 1, but white light in a specific frequency band obtained through a narrow band filter from a white light source is directed toward the objective lens by a beam splitter, and the white light that has passed through the objective lens. An optical system is provided that combines again the reference light obtained by reflecting the light to the reference surface and the measurement light obtained by reflecting the light to the surface of the object to be measured, and forms an image on the imaging surface of the CCD camera. Then, while detecting the optical interference signal by the CCD camera, the position of the PZT where the amplitude of the optical interference signal is maximized by moving the objective lens in the irradiation direction (vertical direction) of the white light by the driving means such as PZT. Becomes the surface of the thin film, and the film thickness of the object is obtained from the amount of vertical movement of PZT.
JP 2001-66122 A

  In the measurement apparatus using the white interferometer described above, the surface shape of the thin film as the object can be known from the position of the PZT where the amplitude of the interference signal is maximized, and the film thickness of the object is determined according to the scanning distance of the PZT. I can know. However, in the case of highly accurate measurement, it is necessary to scan the objective lens correctly by driving means such as PZT, and the mechanical error factor by the driving means cannot be completely eliminated.

  Therefore, in view of the above problems, the present invention eliminates mechanical error factors, and allows the shape of each reflecting surface of the object and the distribution of film thickness to be accurately measured on a secondary plane. It is an object of the present invention to provide a measuring method and a thin film shape measuring apparatus.

The thin film shape measuring method according to claim 1 includes a first step of generating a sine wave-scanned light having an angular frequency ω _b and a plurality of reflecting surfaces which are measurement surfaces of an object by dividing the light. And a second step of combining the object light from the measurement surface of the object and the reference light from the reference surface to obtain interference light after being reflected by the reference surface, the reference surface or the reference surface A third step of sine-wave oscillating the measurement surface of the object at an angular frequency ω _c , and a fourth step of capturing the interference light at two-dimensionally arranged detection points and converting them into electrical interference signals, respectively. A process, a value of the processing signal obtained from the interference signal electrically converted in the fourth step based on a processing signal obtained by Fourier transforming the interference signal, and the process theoretically derived The interference signal changes when the sum of squares of the difference from the signal value is minimized. A fifth step of estimating amplitude Z _bi and phase _{α i (i = 1,2, ...} n) the value of each said detection point, the value of the modulation amplitude Z _bi and phase alpha _i, the subject matter A sixth step of calculating, for each detection point, the optical path difference L _zi , Lα _i between the light reflected by each of the reflecting surfaces and the reference light reflected by the reference surface, and the optical path differences L _zi , Lα and a seventh step of calculating the position of each reflection surface of the object for each detection point based on the value of _i .

In this case, the phase of the electrically converted interference signal changes in a sine wave shape, and the modulation amplitude Z _bi is proportional to the optical path difference L _zi and the scanning amplitude b of the light wave scanned in the sine wave _shape, and the interference The phase α _i that is the time average of the phase of the signal is proportional to the optical path difference Lα _i . Therefore, a modulation amplitude Z optical path difference over the wavelength obtained from the _bi L _zi, by combining the phase α optical path difference equal to or smaller than the wavelength obtained from _i L [alpha _i, the value of the following wavelengths or more of the optical path difference L _i Wavelength It can be calculated with the same high accuracy as the measurement accuracy of the optical path difference Lα _i .

Further, the sum of squares of the difference between the value of the processed signal obtained from the electrically converted interference signal and the value of the processed signal, which is a theoretical formula, is defined as an error function H, and the error function H is minimized. The modulation amplitude Z _bi and the phase α _i for each reflecting surface of the object included in the theoretical formula can be estimated. From the estimated values of the modulation amplitude Z _bi and the phase α _i , the positions of the respective reflecting surfaces of the object are obtained, and the mechanical error factors as in the prior art are eliminated, and the shapes of the reflecting surfaces of the object are determined. Can be measured accurately on the secondary plane.

Such an effect is obtained by a light source device that generates light that is wavelength-scanned in a sine wave form at an angular frequency ω _b and a plurality of reflections that are measurement surfaces of an object by dividing the light from the light source device. An interference optical system that obtains interference light by combining the object light from the measurement surface of the object and the reference light from the reference surface after being reflected by the surface and the reference surface, and the reference surface or the reference surface a sinusoidal oscillation means for sinusoidally vibrating the measurement surface of the object at angular frequency omega _c, the interference light captured at the detection point arranged two-dimensionally, each photoelectric conversion means for converting the electrical interference signals And the value of the processing signal obtained from the interference signal electrically converted by the photoelectric conversion means based on the processing signal obtained by Fourier transforming the interference signal, and the theoretically derived processing signal Before the sum of squares of the difference from the value is minimized Modulation amplitude of the interference signal Z _bi phase _{α i (i = 1,2, ...} n) first calculating means for estimating the value of each said detection point, the value of the modulation amplitude Z _bi and phase alpha _i Second calculation means for calculating, for each detection point, optical path differences L _zi and Lα _i between the light reflected by each reflecting surface of the object and the reference light reflected by the reference surface; 5. A thin film shape measuring apparatus according to claim 4, further comprising: a third calculation unit that calculates the position of each reflecting surface of the object for each detection point based on the values of the optical path differences L _zi and Lα _i. it can.

In the thin film shape measuring method according to claim 2, in the fifth step, the frequency component ω is in a region where ω _c / 2 <ω <3ω _c / 2 and a region where 3ω _c / 2 <ω <5ω _c / 2. , To obtain the following two processed signals A _s (t) and A _c (t) by Fourier transforming the interference signal,

  The sum of squares of the difference between the value of the processed signal obtained from the interference signal converted in the fourth step and the value of the processed signal that is theoretically derived (distinguishes by adding a sign “^”) As an error function H such as

It is characterized in that the values of the modulation amplitude Z _bi and the phase α _i that minimize the value of the error function H are estimated for each detection point.

Thus, the interference signal by Fourier transform, areas in the vicinity of the frequency components _{ω c (ω c / 2 <} ω <3ω c / 2) and a region near _{2ω c (3ω c / 2 <} ω <5ω c / 2), after obtaining the two processed signals A _s (t) and A _c (t), the values of the modulation amplitude Z _bi and the phase α _i that minimize the value of the error function H are estimated. The position of each reflecting surface of the object can be accurately measured.

  Such an effect can also be realized by a thin film shape measuring apparatus having the configuration of claim 5.

The thin film shape measuring method according to claim 3 further includes an eighth step of calculating, for each detection point, a thickness between the reflecting surfaces of the object based on the value of the optical path difference Lα _i L _zi . .

  As a result, not only the shape of each reflecting surface of the object but also the thickness between the reflecting surfaces can be accurately measured on the secondary plane while eliminating mechanical error factors.

  Such an effect can also be realized by a thin film shape measuring apparatus having the configuration of claim 6.

  According to the thin film shape measuring method of claim 1 and the thin film shape measuring apparatus of claim 4, the mechanical error factor is eliminated and the shape of each reflecting surface of the object is accurately measured on the secondary plane. can do.

  According to the thin film shape measuring method of claim 2 and the thin film shape measuring apparatus of claim 5, it is possible to accurately measure the position of each reflecting surface of the object.

  According to the thin film shape measuring method of claim 3 and the thin film shape measuring apparatus of claim 6, the mechanical error factor is eliminated, and not only the shape of each reflecting surface of the object but also the film thickness between the reflecting surfaces. Can be accurately measured on the secondary plane.

Hereinafter, preferred embodiments of a thin film shape measuring method and a thin film shape measuring apparatus according to the present invention will be described in detail with reference to the accompanying drawings. In FIG. 1 showing the overall configuration of the apparatus, in this embodiment, as an interference optical system for realizing the double sine wave phase modulation interferometry, a SWS (Sinusoidal Wavelength-Scanning) light source apparatus 1 and this For example, a Michelson-type laser interferometer 2 that takes in output light from the SWS light source device 1 is provided. The SWS light source device 1 generates light that has been wavelength-scanned in a sine wave form. Specifically, the semiconductor laser 10 as a light source, the lens 11 that makes emitted light from the semiconductor laser 10 parallel light, and the parallel light. A diffraction grating 12 on which light is incident, a mirror 13 for reflecting the first-order diffracted light from the diffraction grating 12 back to the semiconductor laser 10, and a vibration means for rotating and vibrating the mirror 13 in a sinusoidal shape at an angular frequency ω _b And a laser scanner 31. Then, by forming an external resonator using the first-order diffracted light from the diffraction grating 12 between the reflection surface (not shown) on the rear side of the semiconductor laser 10 and the mirror 13, the 0th-order diffracted light of the diffraction grating 12 is The light is output from the SWS light source device 1 as incident light to the laser interferometer 2. In addition, you may use things other than the semiconductor laser 10 as a light source, and the structure of SWS light source device 1 is not limited to the thing in an Example.

The laser interferometer 2 includes a beam splitter 21, a mirror 20 including a piezoelectric element 41 that is a vibrating means, and lenses 22 and 23. The beam splitter 21 divides the output light that has been wavelength-scanned in a sine wave form from the SWS light source device 1 into a mirror 20 that forms a reference surface and a thin film-like object 50, and the reflection from these mirrors 20. The reference light, which is light, and the object light, which is the reflected light from the object 50, are combined to obtain interference light. The piezoelectric element 41 causes the mirror 20 to vibrate in a sine wave at an angular frequency ω _c , whereby the phase of the reference light from the mirror 20 is sine wave phase modulated. The object light from the object 50 consists of light reflected by the front surface and the back surface, which are the measurement surfaces of the object 50, and from the beam splitter 21 that combines the object light and the reference light. The interference light passes through lenses 22 and 23 and is guided onto a detection surface of a CCD image sensor 61 described later. As the laser interferometer 2, various interference optical systems can be used.

  61 is a CCD image sensor corresponding to a two-dimensional photoelectric conversion means, which has a detection surface arranged on the image surface of the interference light formed through the lenses 22 and 23, and the light intensity distribution due to this interference is measured for each pixel. Each is converted into an electrical detection signal and output. Since the photoelectric conversion means here should just be able to measure the light intensity which changes with time in a short time, it is preferable to use a photoelectric conversion element with good follow-up like a CCD (Charge Coupled Device). 62 calculates, based on the detection signal from the CCD image sensor 61, each optical path difference between the object light reflected by the front and back surfaces of the object 50 and the reference light reflected by the mirror 20, The personal computer is a processing apparatus that determines the positions of the front and back surfaces of the object 50 and calculates the thickness of the object 50 from the optical path differences. The personal computer 62 is connected to a display (not shown) such as a monitor for displaying a photographed image of the CCD image sensor 61 as necessary.

  Next, the operation principle of the above configuration will be described. When a predetermined injection current is applied to the semiconductor laser 10 of the SWS light source device 1, light from the semiconductor laser 10 is output toward the lens 11, and parallel light that has passed through the lens 11 enters the diffraction grating 12 and is diffracted. The The first-order diffracted light from the diffraction grating 12 is perpendicularly incident on the mirror 13 with the laser scanner 31 attached to the back surface, and the reflected light from the mirror 13 returns to the semiconductor laser 10 through the diffraction grating 12 and the laser 11 again. As a result, an external resonator is formed between the reflection surface (not shown) on the rear side of the semiconductor laser 10 and the mirror 13 inside the SWS light source device 1.

In this state, when the rotationally oscillate sinusoidally mirror 13 at the angular frequency omega _b by the laser scanner 31, 0 wavelength of the diffracted light of the diffraction grating 12 lambda (t) is scanned sinusoidally as: .

Where λ ₀ is the center wavelength of the semiconductor laser 10 and b is the scanning amplitude. The 0th-order diffracted light of the diffraction grating 12 becomes incident light to the laser interferometer 2 and reaches the beam splitter 21 where it is divided into light propagating to the mirror 20 and light propagating to the object to be measured 50. The Since the mirror 20 is vibrated sinusoidally with the amplitude a and the angular frequency ω _c by the piezoelectric element 41, the phase of the reference light, which is reflected light from the mirror 20, is modulated in a sinusoidal shape (sinusoidal phase modulation). On the other hand, object light, which is reflected light from the object 50, is composed of light reflected by the front and back surfaces of the object 50, respectively. The reference light from the mirror 20 and the object light from the object 50 are combined and overlapped by the beam splitter 21 and reach the detection surface of the CCD image sensor 61 as interference light. In particular, here, the light fields of the front and back reflecting surfaces of the object 50 are formed on the CCD image sensor 61 by the lenses 22 and 23.

  The CCD image sensor 61 converts a two-dimensional intensity distribution resulting from interference between object light and reference light into an electrical interference signal. The interference signal is converted into a digital signal by an A / D converter (not shown) and then taken into the personal computer 62, where calculation for obtaining the surface shape and film thickness distribution of the plurality of reflecting surfaces of the object 50 is performed. Processing is performed. As described above, when the SWS light source device 1 generates light that has been wavelength-scanned in a sine wave form, and further oscillates the mirror 20 by the piezoelectric element 41, the interference signal is further subjected to sine wave phase modulation. The time-varying component S (t) included in the interference signal detected by the image sensor 61 can be expressed by the following equation.

Here, a ₁ and a ₂ are the respective reflectances of the front surface and the back surface of the object 50. When the mirror 20 vibrates with acos (ω _c + θ) and the intensity change M (t) of the light from the SWS light source device 1 is taken into consideration, the interference signal S (t) is expressed by the following equation. .

Assuming that the optical path differences representing the positions of the front and back surfaces of the object 50 are L ₁ and L ₂ , the modulation amplitude Z _{bi in the} above equation is Z _bi = 2πbL _i / λ ₀ ² , and the phase α _i Is α _i = 2πL _i / λ ₀ . That is, if the modulation amplitude Z _bi and the phase α _i are calculated for each of the front surface and the back surface of the object 50, the optical path difference is obtained by the known proportionality constants (2πb / λ ₀ ² ) and (2π / λ ₀ ). the L _i can be identified.

Here PC 62, the interference signal by S (t) is the Fourier transform, a region near the frequency component omega is _{ω c (ω c / 2 <} ω <3ω c / 2), area near 2 [omega _c (3 [omega] _{From c} / 2 <ω <5ω _c / 2), the following two processed signals A _s (t) and A _c (t) are respectively obtained.

The personal computer 62 then processes the value of the processed signal at the time t _m = mΔt obtained from the interference signal S (t) detected by the CCD image sensor 61 and the value of the processed signal (sign “^”) as a theoretical expression. A series of the sum of squares of the difference from the difference is used as an error function H, and a multidimensional search is performed so that the value of the error function H is minimized, and each of the modulation amplitude Z _bi and the phase α _i is determined. Get an estimate. The minimum value of the error function H can be calculated in a short time by continuously substituting the value of the processing signal obtained by the theoretical formula by processing on the software of the personal computer 62. The error function H is expressed as follows:

Here, when the value of the optical path difference L _i obtained from the estimated value of the modulation amplitude Z _bi is L _zi and the value of the optical path difference L _i obtained from the estimated value of the phase α _i is Lα _i , the optical path difference L _zi is equal to or greater than the wavelength. a rough value, the value of the optical path difference L [alpha _i value parts: the magnitude wavelength (phase alpha _i is determined in the range of -π~ + π, the value of the optical path difference L [alpha _i is 1-? _0/2 - a + λ becomes _0/2 range). Therefore, if it is less than the measurement error of the optical path difference L _zi is _{λ 0/2 (<λ 0} /2), the personal computer 62 the two optical path difference L _zi calculated from the value of L [alpha _i, numeric shown in the following equation m Calculate _ci . The measurement accuracy of the optical path difference L _{zi is, Z bi = 2πbL i / λ} 0 As is apparent from the ^second relational expression is proportional to the scan amplitude b of the light wavelength scanning sinusoidally, the optical path difference L _zi to the measurement errors of less than lambda _0/2 is, SWS light source device 1 to increase the scan amplitude b is required.

PC 62 rounds the decimal point numbers m _ci obtained by the above equation, the order of the wavelength lambda ₀ contained in the optical path difference L _zi (integer value) to calculate the m _i, the optical path difference longer than the wavelength of the formula L _i can get.

The value of the optical path difference L _i obtained here is obtained by adding an integral multiple of the wavelength λ ₀ to the value (Lα _i ) of the portion having a size equal to or smaller than the wavelength, and therefore the measurement accuracy is the phase α _i. It is equal to the measurement accuracy by the optical path difference Lα _i obtained from the estimated value, and is on the order of several nm.

When the values of the optical path differences L _zi and Lα _i described above are obtained, the personal computer 62 is located at one position of the object 50 at each position on the two-dimensional plane (XY) detected by the CCD image sensor 61. The shape r ₁ of a certain front surface and the shape r ₂ of the back surface that is the other surface can be calculated using the following equations.

Here, n _R is the refractive index of the object 50. Further, the personal computer 62 can easily calculate the value of the thickness d of the object using the following equation. The integer value m here is m = m ₂ −m ₁ .

In this way, the phase of the interference signal obtained through the laser interferometer 2 and the CCD image sensor 31 changes in a sine wave shape, and the modulation amplitude Z _bi is the optical path difference L _zi and the scanning of light wavelength-scanned in a sine wave _shape. In addition to being proportional to the amplitude b, the phase α _i that is the time average of the phases of the interference signals is proportional to the optical path difference Lα _i . Therefore, a modulation amplitude Z optical path difference over the wavelength obtained from the _bi L _zi, by combining the phase α optical path difference equal to or smaller than the wavelength obtained from _i L [alpha _i, the value of the following wavelengths or more of the optical path difference L _i Wavelength It can be calculated with the same high accuracy as the measurement accuracy of the optical path difference Lα _i . In order to combine the optical path differences L _zi and Lα _i , the measurement accuracy of the optical path difference L _zi needs to be higher than half the center wavelength λ ₀ of the wavelength-scanned light. _Since the measurement accuracy of _zi is proportional to the scanning amplitude b, a desired measurement accuracy can be obtained by increasing the scanning amplitude b.

Particularly in this embodiment, the sum of squares of the difference between the value of the processing signal detected by the CCD image sensor 31 and the value of the processing signal, which is a theoretical formula, is defined as an error function H, and this error function H is minimized. The modulation amplitude Z _bi and the phase α _i for each reflecting surface (the front surface and the back surface of the object 50) included in the theoretical formula can be estimated. From the estimated values of the modulation amplitude Z _bi and the phase α _i , the position of each reflecting surface of the thin film can be obtained, the surface shape of the thin film can be obtained accurately, and the thickness between the reflecting surfaces can also be obtained accurately. it can.

Next, an actual measurement example using the thin film shape measuring apparatus of the present embodiment will be presented with reference to FIGS. Here, quartz glass having a refractive index n _R of 1.46 and a thickness of 20 μm is used as the object 50. The center wavelength λ ₀ of the semiconductor laser 10 used as the SWS light source device 1 is 780 nm, the output is 40 mW, and the laser scanner 31 gives rotational vibration to the mirror 13 so that the wavelength scanning width 2b = 20 nm and the frequency ω _b The light that was wavelength-scanned in a sine wave shape of /2π=66.4 Hz was taken out from the SWS light source device 1. Furthermore, in the laser interferometer 2, the frequency of the sine wave phase modulation of the reference light is ω _c / 2π = 2125 Hz because the piezoelectric element 41 vibrates the mirror 20. As the photoelectric detection means, the two-dimensional CCD image sensor 61 as described above is used, but the number of measurement points (number of pixels) on the XY plane here is 60 × 60, and the measurement area is 1.2. × 1.2 mm. Therefore, the size of each pixel is 20 μm square.

FIG. 2 separately shows the value of the processing signal A _s (t) detected by one pixel of the CCD image sensor 61 and the value of the processing signal theoretically estimated at the same position. . Here, the reflectance a ₁ of the surface of the object 50 and the rough values of the modulation amplitudes Z _b1 and Z _b2 are obtained from the detected value of the processed signal A _s (t), and each phase α ₁ , The initial value of α ₂ was set at an interval of 1.0 rad. The following table shows the initial values of a ₁ , Z _b1 , Z _b2 , α ₁ , α ₂ , H and the values of a ₁ , Z _b1 , Z _b2 , α ₁ , α ₂ , H that minimize the error function H. It shows an estimated value.

In this way, the personal computer 62 estimates each value of the reflectance a _i , the modulation amplitude Z _bi, and the phase α _i for each pixel of the CCD image sensor 61. The following table shows the optical path differences L _z1 , L _z2 , Lα ₁ , Lα ₂ calculated from the above equations, where the coordinate position of each pixel on the XY plane is (I _x , I _Y ), The numerical values m _c1 and m _c2 before rounding are shown, and here, the values measured in units of 10 on the Y axis along the 10th pixel (I _x = 10) on the _x axis are shown. .

At each pixel of the CCD image sensor 61, phase alpha _1, the optical path difference L [alpha ₁ obtained from the estimated value of alpha _2, if L [alpha ₂ is calculated, the surface of the object 50 on a two-dimensional plane from the above equation The shape r ₁ and the shape r ₂ of the back surface can be obtained respectively. FIG. 3 shows an example of the measurement result three-dimensionally. Here, the surface shape and the back surface shape of the object 50 are obtained with repeated measurement errors of about 10 nm or less. Each pixel size (ΔI _x , ΔI _Y ) along the XY axis is 20 μm as described above.

FIG. 4 three-dimensionally shows the distribution of the orders m ₁ and m ₂ obtained by rounding off the decimal points of the numerical values m _c1 and m _c2 for each pixel. For the order m ₁ corresponding to the surface of the object 50, here m ₁ = 30 occupies 16% of the whole, m ₁ = 31 occupies 73% of the whole, and m ₁ = 32 occupies 11% of the whole. Therefore, the personal computer 62 adopted the value of m ₁ = 31, which has the highest occupation ratio. As for the order m ₂ , m ₂ = 101 accounted for 4% of the total, m ₂ = 102 accounted for 80% of the total, and m ₂ = 103 accounted for 16% of the total. The value of m ₂ = 102, which has the highest occupation rate, was adopted. Therefore, the value of m = m ₂ −m ₁ in this case is 71.

If the surface shape r ₁ and the back surface shape r _{2 of the} object 50 and the integer value m are determined, the thickness between the surface and the back surface of the object 50 can be measured from the above equation. FIG. 5 shows the measurement results three-dimensionally. The thickness of the object 50 can also be calculated with a repeated measurement error of about 10 nm or less.

In the above-described method of rotating and vibrating the mirror 13 in a sine wave shape, the SWS light source device 1 provides light that has been accurately wavelength-scanned in a sine wave shape with a wide scanning width of 2b = 20 nm. If the interference signal generated from the wavelength scanned light is detected by the CCD image sensor 31 developed two-dimensionally, each value of the modulation amplitude Z _bi and the phase α _i with respect to the _front and back surfaces of the object 50 is obtained. processing by the processing of the personal computer 62 signals a _s (t), it is possible to estimate from the a _c (t), can be determined integer value m _i. As a result, the optical path difference L _i greater than the wavelength can be obtained by the measurement error due to the optical path difference Lα _i obtained from the estimated value of the phase α _i , and the surface shape r ₁ and the back surface of the object 50 can be obtained with an error of 10 nm. The shape r ₂ can be measured simultaneously, and the thickness d of the object 50 can also be measured.

  In the above embodiment, it is assumed that the object 50 is a single-layer thin film. However, in the case of a multilayer film having a plurality of reflecting surfaces, i in the above equation is also set to 3 or more (i = 1, 2,... N), the shape of each reflection surface of the object 50 and the film thickness between the reflection surfaces can be individually measured. As the object 50, a thin film on a semiconductor wafer, a plastic film, a coating film, or the like can be applied.

As described above, in the present embodiment, the first step of generating a sine wave-scanned light having an angular frequency ω _b and a plurality of reflecting surfaces which are the measurement surfaces of the object 50 by dividing the light. (Front surface, back surface) and the second step of obtaining interference light by combining the object light from the measurement surface of the object 50 and the reference light from the mirror 20 after reflection on the reference surface by the mirror 20 And the third step of vibrating the reference surface or the measurement surface of the object 50 at the angular frequency ω _c and the detection point of the CCD image sensor 61 in which the interference light is arranged two-dimensionally, On the basis of a fourth step of converting the interference signal S (t) to a typical interference signal S (t) and the processed signals A _s (t) and A _c (t) obtained by Fourier transforming the interference signal S (t). the value of the electrically converted the resulting from the interference signal processing signal _{a s (t), a c} (t) in step, theoretical When the square sum of the difference between the value of the processed signal derived is minimized, modulation amplitude Z _bi and phase _{α i (i = 1,2, ...} n) contained in the interference signal S (t) of A fifth step of estimating the value for each effective detection point of the CCD image sensor 61, and the light reflected by each reflecting surface of the object 50 from the values of the modulation amplitude Z _bi and the phase α _i ; the optical path difference L _zi and the reference light reflected by the mirror 20, a sixth step of calculating the L [alpha _i for each detection point of the CCD image sensor 61, the optical path difference L _zi, based on the value of L [alpha _i, the And a seventh step of calculating the position of each reflecting surface of the object 50 for each detection point of the CCD image sensor 61.

  In this embodiment, the SWS light source device 1 as a light source device that performs the first step, the laser interferometer 2 as the interference optical system that performs the second step, and the sine wave vibration means that performs the third step. Piezoelectric element 41, a CCD image sensor 61 as photoelectric conversion means for performing the fourth step, first calculation means for performing the fifth step, second calculation means for performing the sixth step, and seventh A device including a personal computer 62 as a third calculation means for performing the process is employed.

In the above method and apparatus, the phase of the electrically converted interference signal S (t) changes sinusoidally, and the modulation amplitude Z _bi is the optical path difference L _zi and the scanning amplitude of the light that has been wavelength-scanned sinusoidally. In addition to being proportional to b, the phase α _i, which is the time average of the phase of the interference signal S (t), is proportional to the optical path difference Lα _i . Therefore, a modulation amplitude Z optical path difference over the wavelength obtained from the _bi L _zi, by combining the phase α optical path difference equal to or smaller than the wavelength obtained from _i L [alpha _i, the value of the following wavelengths or more of the optical path difference L _i Wavelength It can be calculated with the same high accuracy as the measurement accuracy of the optical path difference Lα _i .

Further, the sum of squares of the difference between the value of the processed signal obtained from the electrically converted interference signal S (t) and the value of the processed signal as a theoretical formula is defined as an error function H, and this error function H is minimized. By doing so, it is possible to estimate the modulation amplitude Z _bi and the phase α _i for each reflecting surface of the object 50 included in the theoretical formula. From the estimated values of the modulation amplitude Z _bi and the phase α _i , the position of each reflecting surface of the object 50 is obtained, and the mechanical error factors as in the prior art are eliminated, and each reflecting surface of the object 50 is removed. Can be accurately measured on the secondary plane.

In the present embodiment, in the fifth step, the interference signal is obtained in a region where the frequency component ω is ω _c / 2 <ω <3ω _c / 2 and a region where 3ω _c / 2 <ω <5ω _c / 2. The two processing signals A _s (t) and A _c (t) shown in Equation 10 above are obtained by performing Fourier transform, and the processing obtained from the interference signal S (t) converted in the fourth step. The sum of squares of the difference between the signal value and the theoretically derived value of the processed signal is defined as an error function H as shown in Equation 11, and the modulation amplitude Z _{bi at} which the value of the error function H is minimized. And the phase α _i are estimated for each detection point of the CCD image sensor 61.

In the present embodiment, the first computing means included in the personal computer 62 is configured such that the frequency component ω is a region where ω _c / 2 <ω <3ω _c / 2 and a region where 3ω _c / 2 <ω <5ω _c / 2. Then, the interference signal S (t) obtained by Fourier transforming the interference signal to obtain the two processed signals A _s (t) and A _c (t) shown in the above equation 10 and converted by the CCD image sensor 61. The sum of squares of the difference between the value of the processed signal obtained from the above and the theoretically derived value of the processed signal is defined as the error function H as shown in Equation 11, and the value of the error function H is minimized. The modulation amplitude Z _bi and the phase α _i are estimated for each detection point of the CCD image sensor 61.

According to the above method and apparatus, the interference signal S with Fourier transform (t), a region _{(ω c / 2 <ω <} 3ω c / 2) in the vicinity of the frequency component omega _c, the region near 2 [omega _c ( After obtaining two processed signals A _s (t) and A _c (t) from 3ω _c / 2 <ω <5ω _c / 2), the modulation amplitude Z _bi and the phase at which the value of the error function H is minimized are obtained. If the value of α _i is estimated, the position of each reflecting surface of the object 50 can be accurately measured.

In this embodiment, the method further includes an eighth step of calculating the thickness between the reflecting surfaces of the object 50 for each detection point of the CCD image sensor 61 based on the value of the optical path difference Lα _i L _zi. Adopted.

In this embodiment, the fourth computing means for calculating the thickness between the reflecting surfaces of the object 50 for each detection point of the CCD image sensor 61 based on the values of the optical path differences L _zi and Lα _i is a personal computer. 62 is further equipped.

  According to the above method and apparatus, not only the shape of each reflecting surface of the object 50 but also the thickness between the reflecting surfaces is accurately measured on the secondary plane while eliminating mechanical error factors. It becomes possible.

  In addition, this invention is not limited to said each Example, A various deformation | transformation implementation is possible. For example, in the present embodiment, the configuration in which the mirror 20 serving as the reference surface is vibrated is shown, but the measurement surface of the object 50 may be vibrated. Furthermore, the SWS light source device 1 is applicable not only to the light source of the SWS interferometer according to the present invention but also to other uses.

It is a schematic explanatory drawing which shows the whole structure of the thin film shape measuring apparatus in one preferable Example of this invention. FIG. 6 is a graph showing separately the value of the processing signal detected by one pixel of the CCD image sensor and the value of the processing signal theoretically estimated at the same position. It is the graph which showed the shape of the surface of a to-be-targeted object, and the shape of a back surface separately, respectively. FIG. 6 is a graph showing separately the distributions of orders m ₁ and m ₂ of the object. It is a graph which shows the thickness of a target object same as the above.

Explanation of symbols

1 SWS light source device (light source device)
2 Laser interferometer (interference optical system)
41 Piezoelectric element (sinusoidal vibration means)
50 Object
61 CCD image sensor (photoelectric conversion means)
62 Personal computer (first computing means, second computing means, third computing means, fourth computing means)

Claims (6)

A first step of generating a sinusoidally wavelength scanned light at an angular frequency ω _b ;
After dividing the light and reflecting each of the plurality of reflection surfaces and the reference surface, which are measurement surfaces of the object, object light from the measurement surface of the object and reference light from the reference surface A second step of combining to obtain interference light;
A third step of sine-wave oscillating the reference surface or the measurement surface of the object with an angular frequency ω _c ;
A fourth step of capturing the interference light at detection points arranged two-dimensionally and converting the interference light into an electrical interference signal;
Based on the processed signal obtained by Fourier transforming the interference signal, the value of the processed signal obtained from the interference signal electrically converted in the fourth step and the value of the processed signal theoretically derived A fifth step of estimating the value of the modulation amplitude Z _bi and the phase α _i (i = 1, 2,... N) of the interference signal for each detection point when the sum of squares of the difference between the two is minimized. ,
Based on the values of the modulation amplitude Z _bi and the phase α _i , optical path differences L _zi and Lα _i between the light reflected by each reflection surface of the object and the reference light reflected by the reference surface are determined as the detection points. A sixth step of calculating each time;
A seventh step of calculating the position of each reflecting surface of the object for each detection point based on the values of the optical path differences L _zi and Lα _i ;
A thin film shape measuring method comprising: In the fifth step, Fourier transform is performed on the interference signal in a region where the frequency component ω is ω _c / 2 <ω <3ω _c / 2 and a region where 3ω _c / 2 <ω <5ω _c / 2. , Obtain the following two processed signals A _s (t) and A _c (t),

The sum of squares of the difference between the value of the processed signal obtained from the interference signal converted in the fourth step and the value of the processed signal that is theoretically derived (distinguishes by adding a sign “^”) As an error function H such as

2. The thin film shape measuring method according to claim 1, wherein the value of the modulation amplitude Z _bi and the phase α _{i at} which the value of the error function H is minimized is estimated for each detection point. 3. The method according to claim 1, further comprising an eighth step of calculating, for each of the detection points, a thickness between the reflection surfaces of the object based on the value of the optical path difference Lα _i L _zi . Thin film shape measurement method. A light source device for generating light that is wavelength-scanned in a sinusoidal form at an angular frequency ω _b ,
After the light from the light source device is divided and reflected by a plurality of reflection surfaces and reference surfaces, which are measurement surfaces of the object, object light from the measurement surface of the object, and from the reference surface An interference optical system that obtains interference light by combining with reference light;
A sinusoidal oscillation means for sinusoidally oscillating at an angular frequency omega _c the measurement surface of the reference surface or the subject matter,
Photoelectric conversion means for capturing the interference light at detection points arranged two-dimensionally and converting them into electrical interference signals, and
Based on the processed signal obtained by Fourier transforming the interference signal, the value of the processed signal obtained from the interference signal electrically converted by the photoelectric conversion means, and the theoretically derived value of the processed signal First arithmetic means for estimating the value of the modulation amplitude Z _bi and the phase α _i (i = 1, 2,... N) of the interference signal when the sum of squares of the difference between the two is minimized. ,
From the values of the modulation amplitude Z _bi and the phase α _i , optical path differences L _zi and Lα _i between the light reflected by each reflection surface of the object and the reference light reflected by the reference surface are determined as the detection points. A second calculation means for calculating each time;
Third arithmetic means for calculating the position of each reflecting surface of the object for each detection point based on the values of the optical path differences L _zi and Lα _i ;
A thin-film shape measuring apparatus comprising: The first calculation means performs Fourier transform on the interference signal in a region where the frequency component ω is ω _c / 2 <ω <3ω _c / 2 and a region where 3ω _c / 2 <ω <5ω _c / 2. To obtain the following two processed signals A _s (t) and A _c (t) respectively,

The square sum of the difference between the value of the processing signal obtained from the interference signal converted by the photoelectric conversion means and the value of the processing signal that is theoretically derived (distinguishes by adding a sign “^”) is as follows: As an error function H like the equation,

5. The thin film shape measuring apparatus according to claim 4, wherein the value of the modulation amplitude Z _bi and the phase α _{i at} which the value of the error function H is minimized is estimated for each detection point. 6. The apparatus according to claim 4, further comprising fourth calculation means for calculating a thickness between reflection surfaces of the object for each detection point based on the values of the optical path differences L _zi and Lα _i. The thin film shape measuring apparatus described.

Images