JP2005156188A - Method and apparatus for measuring flatness of sample and complex dielectric constant by measuring transmission of light - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To realize an optical system capable of measuring both of the flatness of a substrate and the complex dielectric constant of a thin film by measuring a transmission spectrum of a sample. <P>SOLUTION: The transmission spectrum of the parallel flat plate-shaped substrate does not depend on an incidence angle in the peak frequency of a fringe and transmissivity is constant to take the maximum value but, when the incident angle is increased, the transmissivity approaches zero in the peripheral frequency of the fringe. When the thin film is placed on the substrate to increase the thickness of the substrate, the peak frequency of the fringe is shifted to a low-frequency side. Because of these three effects, the spectrum of the ratio of the transmission spectrum of the system composed of the substrate and the thin film to the transmission spectrum only of the substrate at a high incidence angle becomes the spectrum of a structure that the maximum and minimum values are adjacent and the complex dielectric constant of the thin film is calculated. Further, if the structure described above appears in the spectrum of the ratio with the transmission spectrum of a reference position by performing the same measurement while moving a light irradiation position on the substrate, the flatness of the substrate is calculated from the spectrum. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an apparatus for measuring the flatness and complex permittivity of a substrate and a thin film on the substrate by measuring the transmission spectrum of the substrate and the thin film on the substrate, and a measuring method thereof.

  By the capacitance method evolved from the capacitance measurement of capacitors, the complex dielectric constant of a thin film on a substrate can be measured using an LCR meter below several GHz (for example, see “Patent Document 1” below). The measurement limit on the high frequency side in the capacitance method is because it is difficult to correct the effect of electrode loss and the effect of LC resonance due to electrode inductance.

  In the complex permittivity measurement at a high frequency, the resonator method is generally used. The complex dielectric constant of a thin film on a substrate can be obtained from the measurement of changes in intensity and phase with respect to the propagation direction using a network analyzer in a stripline or a microstripline composed of the thin film and electrodes (for example, “Patents below” Reference 2 ”). This method enables complex permittivity measurement in the 0.1 GHz to 10 GHz region. The measurement limit on the high frequency side here is because it is difficult to completely separate and remove the loss of the conductor constituting the line to obtain the characteristics of only the thin film.

  Furthermore, in order to measure the complex dielectric constant of the thin film on the substrate in the high frequency region, the cavity resonator method is used. The resonant frequency and Q value when the thin film sample on the substrate is inserted into the cavity resonator are measured with a network analyzer to obtain the complex dielectric constant of the thin film (for example, see “Patent Document 3” below). ). This method makes it possible to measure the complex dielectric constant in the 1 GHz to 50 GHz region. The measurement limit on the high frequency side here is determined by the physical size of the cavity resonator. In other words, the size of the cavity resonator is about the wavelength (about 6 mm at 50 GHz), and at this time, the dimensional accuracy of the cavity resonator almost coincides with the machining accuracy of manufacturing the resonator. If the dimensional accuracy is low, a large measurement error occurs.

  The above-described capacitance method and resonator method are destructive measurement methods that require processing on the measurement sample to be inserted in order to match the dimensions of the measuring instrument, and processing the sample requires considerable labor and cost. In addition, if the outer dimensions of the manufactured sample, particularly the dimensional accuracy of the portion in contact with the inner wall of the measuring instrument, is low, a large measurement error occurs and accurate measurement becomes difficult.

  There are two non-destructive measurement methods that do not add processing to the sample being measured. One is a method of sandwiching a sample between waveguides, and the other is a method of irradiating the sample with light.

  A sample is sandwiched between two waveguides, the reflection coefficient of one aperture and the transmission coefficient of the other aperture are measured with a network analyzer, and their absolute values and phase angles are solved by solving the Maxwell equation. By substituting into the derived simultaneous equations, the complex dielectric constant of the thin film on the substrate is obtained (for example, see “Patent Document 4” below). This method is called a non-resonator method and is a non-destructive measurement. This method enables complex permittivity measurement in the 1 GHz to 100 GHz (wavelength of about 3 mm) region. The measurement limit on the high frequency side here is due to the work accuracy of the waveguide, as in the above paragraphs [0004] and [0005].

  For complex permittivity measurement, apart from the “electrical measurement method” from the low frequency side to the high frequency side in paragraphs [0002] to [0007], the “optical measurement method” from the high frequency side to the low frequency side. There is. In general, the optical measurement technique can measure a complex dielectric constant under non-destructive, non-contact and atmospheric pressure. These optical measurement methods are called free space methods.

  In the method of calculating the complex permittivity from the optical response (reflected light or transmitted light) at the time when the sample is irradiated with light, the ratio of d / λ is given by assuming that the thickness of the sample is d and the measurement wavelength is λ. The smaller the value, the more difficult it is to measure the complex dielectric constant. This is because light is a wave, and it advances while repeating the “mountain” and “valley” of the wave. When the sample becomes thin, for example, when d / λ = 0.001, the sample interacts with only a small portion between the “crest” and “valley” of the incident light, and a DC electric field appears to act from the sample. Looks like. In DC, the capacitor is just an insulator, and the complex dielectric constant approaches a real constant. For the above reasons, it is difficult to measure the complex dielectric constant when the ratio d / λ is small.

When light is incident on a non-opaque sample and its transmission spectrum is measured to determine the complex dielectric constant of the thin film, the measurement becomes difficult when the thin film is thin and the measurement wavelength is long. The results of this situation calculation are shown in FIG. Here, the thickness of the sample and the complex dielectric constant are fixed, and the result is a calculation result of the transmittance spectrum when the measurement wavelength is changed. The thickness and refractive index of the substrate (S) are 692 μm and 3.4155, respectively, and the thickness and refractive index of the thin film (F) are 0.4 μm and 1.812, respectively. In each figure, the solid line is the transmittance spectrum (T (F / S)) of the sample consisting of a thin film on the substrate, and the dotted line is the transmittance spectrum (T (S)) of the substrate only. In these figures, the spectrum is Only about 2 fringes are drawn. FIG. 2A shows the case where the incident wavelength is around 5 μm (mid-infrared light) and d / λ = 0.08, and FIGS. 2B to 2D show the cases where the wavelength is increased by one digit. In FIG. 2D, the wavelength is around 5 mm (frequency is around 60 GHz) and d / λ = 0.00008. 2 (c) and 2 (d), there is almost no difference between the transmittance spectrum of the substrate alone (T (S)) and the transmittance spectrum of the thin film on the substrate (T (F / S)). It is very difficult to obtain the complex dielectric constant of the thin film. Here, a relationship of “the square of n is equal to ε (n ² = ε)” is established between the complex dielectric constant (ε) and the complex refractive index (n). When simply referring to the refractive index, it means the real part of the complex refractive index.

  The free space method includes a method of directly measuring the amplitude and phase of the reflection coefficient at a fixed incident angle (see, for example, “Non-Patent Document 1” below), the dependence of the reflectance on the angle of incidence, the dependence of the reflectance on the sample thickness, There is a method for obtaining the complex dielectric constant from the frequency dependence of the reflectance (for example, see “Non-Patent Document 2” below). In the measurement method at a fixed incident angle, an expensive measuring instrument such as a network analyzer is required to measure the amplitude and phase of the reflection coefficient. In the method of changing the incident angle, only a measurement of energy reflectance is performed, so that a network analyzer is not necessary. However, in common with both measurements, the absolute value of the reflectance must be measured. For this purpose, a metal flat plate having the same size as that of the sample needs to be used as a reference sample, and there is a problem that a measurement error occurs unless the size and installation position of the metal flat plate are the same as those of the sample to be measured.

  New reflectometry methods have been developed that do not require a metal reference sample. This is a method for obtaining the complex dielectric constant from the ratio of the TE wave reflection coefficient and the TM wave reflection coefficient of the sample when the sample is irradiated with a circularly polarized electromagnetic wave in the millimeter wave band. (See Patent Document 5).

The sensitivity of the measurement in the free space method described in paragraphs [0011] and [0012] is generally low. Top data so far shows that Brewster irradiates a 925 GHz (λ = 324 μm) submillimeter wave to a low dielectric (Low-k) polymer thin film with a thickness of d = 3.27 μm on a silicon substrate while changing the incident angle. The complex dielectric constant is determined by measuring the reflectance before and after the angle (see, for example, “Patent Document 3” below). At this time, the ratio of d / λ is 0.01.
It is.

  As described in paragraphs [0002] to [0007], starting from an electrical measurement method, as described in paragraphs [0008] to [0013], starting from an optical measurement method, Measurement of the complex dielectric constant of a thin film on a substrate between 30 GHz and 3 THz in frequency (100 μm to 10 mm in wavelength) is generally a difficult measurement.

  The frequency bands currently used in communication systems include the 1.9 GHz band and the 2.45 GHz band quasi-microwave band, and the 19 GHz band quasi-millimeter wave band. The quasi-microwave band is allocated to private wireless devices such as personal handy phone systems (PHS) and medium-speed wireless LANs. On the other hand, the quasi-millimeter wave band is allocated to the local wireless device of the high-speed wireless LAN.

  Further, the high frequency region of 30 GHz to 3 THz is a region where future development is expected. Research and development of cordless communication systems in the 50 GHz band, in-vehicle radars for collision prevention in the 60 GHz band, and ultra-high-speed wireless LAN are thriving, and a great leap in information and communication technology is expected. Furthermore, the high-frequency region is currently being put into practical use in millimeter-wave and submillimeter-wave astronomy and fusion plasma research and development, and plays an important role. In order to develop a new device for this high frequency region, measurement of the complex dielectric constant of conventional materials and new materials in this high frequency region is indispensable and an important technology.

  With the high integration and miniaturization of devices in the semiconductor industry, quality requirements for semiconductor wafers have become strict. In particular, a semiconductor wafer serving as a substrate is required to have high flatness, and each time, the demand has been met by remarkable progress in polishing technology.

  However, in order to improve the flatness, not only a polishing technique but also a highly accurate flatness measuring method and apparatus for evaluating it are required. Non-contact and widely used methods for measuring the flatness of the entire wafer surface that are widely used for measuring the thickness of commercialized semiconductors are the capacitance method and the optical interference method.

  In the capacitance method, a sample is inserted between two electrodes (flat plate capacitors) facing each other, and a change in capacitance is detected to measure the local thickness of the sample. The condenser is scanned on the surface of the sample to obtain the flatness of the entire surface (see, for example, “Patent Document 6” below). The capacitance method has the advantage of less influence of particles compared to the optical interference type flatness measurement method described in paragraph [0020], and has various wafer thicknesses and flatnesses from sliced wafers to patterned wafers. It can be measured without contact. However, this method requires application of a surfactant aqueous solution to the surface of the semiconductor wafer, and further requires pretreatment such as removal of a natural oxide film present on the surface of the semiconductor wafer.

  In the interference method, for example, infrared light is irradiated on a semiconductor wafer, and reflected light from a sample is converted into an electrical signal by a detector. The measured spectrum shows fringes due to multiple reflection inside the semiconductor wafer. The local thickness of the sample is obtained from the fringe spacing. The flatness of the entire surface is obtained by scanning the light irradiation position on the sample surface (for example, refer to “Patent Document 7” below). This method does not require pretreatment and can also be measured in situ during the polishing process.

  Flatness that can be achieved by a polishing method that is currently in practical use is 1 to 5 μm (see, for example, “Patent Document 7” below). Before the polishing process, the surface shape of the substrate is measured with a flatness measuring device, and the substrate is deformed while being vacuum-adsorbed with a correction chuck based on the measurement result, and then corrected to a desired shape. When polishing is performed while maintaining the shape, a flatness of 0.3 μm can be achieved (see, for example, “Patent Document 8” below).

JP 2002-286771 A JP 11-166952 A JP 2002-228600 A JP 2002-214161 A JP 2000-193608 A JP-A-10-281710 JP-A-8-216061 JP-A-5-315307 Functional materials, Vol. 18, No10, (1998), p.47 IEICE Transactions, B-II, Vol. J80-B-II, No10, (1997), p.906 Applied Physics Letter vol.74, (1999), 2113-2115

  The present invention has been made in view of the state of the prior art, and the future product development direction is on the high frequency side from 30 GHz, and the low-k (low dielectric material; In paragraphs [0026] and [0027], there is a great demand for measurement of the complex dielectric constant of a thin film, and a technique capable of measuring the complex dielectric constant even if the thickness of the thin film is 1 μm or less in this frequency region. For example, since it can be put into practical use as an in-situ product management device at the actual manufacturing site in the semiconductor industry, we aim to develop a technology that enables measurement of the complex dielectric constant of a thin film on a substrate at a frequency higher than 30 GHz. However, since the flatness of semiconductor wafers actually used in the semiconductor industry is larger than 1 μm, the above-mentioned goal cannot be achieved by the development of a conventional high-sensitivity complex dielectric constant measuring apparatus. The present invention provides a method and apparatus capable of measuring both the flatness of a substrate and the complex dielectric constant of a thin film on the substrate with the same measuring apparatus, and measuring the complex dielectric constant even when the thickness of the thin film is 1 μm or less. Technical issue.

で表される。ここで、ｃとＮはそれぞれ光速度と整数で、ν_ｓ、ｄ_ｓ、ｎ_ｓ、θは、それぞれピーク周波数、基板の厚さ、基板の屈折率、入射角度である。同様に、基板上の薄膜の透過スペクトルにもフリンジが現れ、このフリンジのピーク周波数は、

で表される。ここで、ν_ｆ、ｄ_ｆ、ｎ_ｆは、それぞれピーク周波数、薄膜の厚さ、薄膜の屈折率である。 At wavelengths where the semiconductor substrate is transparent, transmittance measurements are possible. Further, if the substrate is a parallel plate, fringes appear in the transmission spectrum due to multiple reflection within the substrate. The fringe peak frequency is

It is represented by Here, c and N are respectively the speed of light and an integer, and ν _s , d _s , n _s , and θ are a peak frequency, a substrate thickness, a substrate refractive index, and an incident angle, respectively. Similarly, a fringe appears in the transmission spectrum of the thin film on the substrate, and the peak frequency of this fringe is

It is represented by Here, ν _f , d _f , and n _f are a peak frequency, a thickness of the thin film, and a refractive index of the thin film, respectively.

である。 The displacement Δν (= ν _f −ν _s ) of the peak frequency in the sample in which a thin film is formed on the substrate from the peak frequency of the substrate is obtained from (Equation 1) and (Equation 2).

It is.

Here, first, the case of a high dielectric (High-k) film will be estimated. For example, a high-k material such as silicon (n _s = 3.4 and d _s = 700 μm) and a thin film (thickness d _f = 1 μm), such as metal, can have n _f ˜100 or more. At this time, the displacement amount of the peak position when the millimeter wave near 65 GHz is irradiated at normal incidence is obtained as −2.7 GHz from (Equation 3). On the other hand, from (Equation 1), the peak interval of the fringe is obtained as 63 GHz. Since the peak position is shifted to the low frequency side by about 4% (= -2.7 / 63) in one fringe with and without the high-k thin film, this measures the transmission spectrum of each sample. Then, it is a detectable amount, and the complex dielectric constant of a high-k thin film having a thickness of 1 μm can be obtained.

Low-k薄膜が有る時と無い時で、ピーク位置が、１フリンジの中で最大約0.078％（＝Δν／ν_ｓ＝−0.04916／63.02520）しか低周波数側へ変位しない。このために各々の試料の透過スペクトルを測定しても、ピーク周波数の変移量を検出できず、Low-k薄膜の複素誘電率を求めることができない。
表１の計算では、シリコンの複素屈折率の実数部が有限の値（ｎ_s=3.4）で、虚数部をゼロ（ｋ＝0）としているので、透過ピークでは透過率は１００％になっている。さらに、表１の第３行目（ν_ｓ）は、フリンジのピーク間隔であるが、（数１）からもわかるように、これらの値は、（数１）で最初に現れるフリンジのピーク位置（N=1）の周波数でもある。 Next, let us estimate the case of a low dielectric (Low-k) film. For example, the substrate is silicon (n _s = 3.4 and d _s = 700 μm), and the thin film (thickness d _f = 1 μm) is n _f = 1.8 in a low-k material such as a silicon thermal oxide film (SiO ₂ ). . Table 1 summarizes the results of calculating the peak frequency displacement Δν (Equation 3) and the fringe peak interval ν _s (Equation 1) when a millimeter wave near 65 GHz is irradiated while changing the incident angle.

With and without the low-k thin film, the peak position is displaced to the low frequency side by a maximum of about 0.078% (= Δν / ν _s = −0.04916 / 63.02520) in one fringe. For this reason, even if the transmission spectrum of each sample is measured, the peak frequency shift cannot be detected, and the complex dielectric constant of the low-k thin film cannot be obtained.
In the calculation of Table 1, since the real part of the complex refractive index of silicon is a finite value ( _ns = 3.4) and the imaginary part is zero (k = 0), the transmittance is 100% at the transmission peak. Yes. Furthermore, the third row (ν _s ) of Table 1 is the fringe peak interval, but as can be seen from (Equation 1), these values are the peak positions of the fringe that appear first in (Equation 1). It is also the frequency of (N = 1).

FIG. 3 shows the incident angle dependence of the transmittance (Ts and Tp) and reflectance (Rs and Rp) of S deflection and P deflection. Here, it is a calculation result when a silicon substrate (n _s = 3.4 and d _s = 700 μm) is irradiated with a 60 GHz millimeter wave. As can be seen from Table 1, the incident frequency of 60 GHz is shifted to the lower frequency side than the peak frequency of the fringe of this sample. Paying attention to the transmittance (Ts) of S deflection, when the incident angle is increased, the transmittance decreases monotonically and becomes zero at 90 degrees. After the transmittance (Tp) of P deflection reaches the maximum value at the Brewster angle near 75 degrees, when the incident angle is increased, the transmittance decreases monotonically and becomes zero at 90 degrees.

  From the paragraph [0027], at the fringe peak frequency that appears due to multiple reflection in the substrate, the transmittance takes the maximum value (the transmittance is 100% at k = 0) regardless of the incident angle. On the other hand, at the frequency deviated from the peak frequency from paragraph [0028], the transmittance approaches zero when the incident angle is increased. When these two effects overlap, the transmission spectrum becomes thinner and narrower as the incident angle is increased. This is shown in FIG.

From the second row (Δν) in Table 1 of paragraph [0027], the peak frequency of the transmittance spectrum is shifted between the substrate and the thin film on the substrate. The transmission spectrum (T (S)) of the substrate and the transmission spectrum (T (F / S)) of the thin film sample on the substrate were measured at a large incident angle (oblique incidence), and the ratio (relative transmittance; T ( Taking F / S) / T (S)), a curve with adjacent maximum and minimum values is obtained due to the effect of paragraph [0029]. This curve has a rotational symmetry center of 180 degrees. This is shown in FIGS. 5 (a) and 5 (b). In this calculation, the substrate is silicon (n _s = 3.4155 and d _s = 700 μm), the thin film is a thermal oxide film of silicon (n _f = 1.812), and FIG. 5 (a) shows the thickness of this thin film being 2 μm. 5 (b) has a thickness of 0.4 μm. In this calculation result, when the incident angle is 85 degrees, a peak of ± 7% appears when the film thickness is 2 μm, and a peak of ± 1.4% appears when the film thickness is 0.4 μm.

  In the calculations so far, the imaginary part (k) of the complex refractive index of both the substrate and the thin film has been set to zero. When the k of the substrate is not zero, the peak of the transmittance spectrum of the substrate and the thin film on the substrate falls below 100%, but since both drop by the same, taking their ratio (see paragraph [0030]) Qualitatively, it is the same as FIGS. 5 (a) and 5 (b). On the other hand, when the k of the thin film is not zero, the peak of the transmittance spectrum of the thin film on the substrate falls below that of the substrate. As a result, when the ratio of the transmittance spectrum of the substrate to the thin film on the substrate is taken, the maximum value is lowered and the rotational symmetry center of 180 degrees is eliminated.

In FIG. 6 of the measurement result of the sample covered with a thin film in the shape of a half moon in the following paragraph [0035], the complex dielectric constant of the thin film is obtained by selecting the complex dielectric constant of the thin film so as to best fit the relative transmittance curve. The real part ε ₁ = 3.5 and the imaginary part ε ₂ = 0.08. In FIG. 8 of the measurement result of the silicon substrate sample having a uniform thickness in the following paragraph [0037], it can be estimated that the 4-inch silicon substrate has a substantially wedge shape when measured with a millimeter wave near 65 GHz. Next, when the mirror symmetry angle and the flatness of the substrate are obtained so as to best fit these relative transmittance curves, the angle is 11.2 degrees and the flatness is about 20 μm.
In this way, both the flatness of the substrate and the complex dielectric constant of the thin film on the substrate are measured with the same measuring device by optical measurement regardless of electrical measurement, and even when the thickness of the thin film is 1 μm or less, The dielectric constant can be measured.

  The best mode for carrying out the invention will be described below.

Embodiments of complex permittivity measurement according to the present invention will be described with reference to the drawings.
FIG. 1 is a layout diagram of the complex permittivity measuring apparatus 10. The CW light emitted from the light source 12 (in this figure, the millimeter wave backward wave tube (BWO)) is subjected to intensity modulation by the mechanical chopper 14. The light that has passed through the lens 15 and the aperture 16 becomes a plane wave. Light is condensed on the sample surface by the lens 17 and the aperture 18 on the front surface of the sample 11. A polarizer (not shown in this figure) and an optical power attenuator (not shown in this figure) are inserted into the incident system 30 as necessary. Only the light transmitted through the sample is received by the lens 19 and the aperture 20 to be converted into a plane wave. This light is received by the lens 21 and the aperture 22 and condensed on a detector (Golay cell in this figure). The light intensity signal is converted into an electric signal by a detector and sent to a measuring instrument (not shown in this figure). From the sample to the detector is referred to as a light receiving system 31. A light source, a sample, a detector, and the like are arranged substantially in a straight line. The traveling direction of this light is taken on the z axis. The light source is mounted on an xy automatic stage (not shown in this figure) in order to arbitrarily change the light incident position on the sample. In order to change the incident angle, the sample is placed on an automatic rotation stage (not shown in this figure) and can freely rotate around the vertical axis (y-axis). The detector is placed on an xyz automatic stage (not shown in this figure) and an automatic rotation stage (not shown in this figure) so that it can be installed at an optimum position. The sample holder (not shown in this figure) has been devised so as not to block incident light even at oblique incidence. Furthermore, in order to prevent light that does not pass through the sample from entering the light receiving system, the sample holder has a radio wave absorber. (Omitted in this figure) is attached. Lenses 17 and 19 are not used when parallel light is incident on the sample.

A 10 μm thick thermal oxide film (SiO ₂ ) is deposited on both sides of a 4 inch diameter silicon substrate with a thickness of 700 μm. A sample with the exposed silicon surface was prepared. This sample was set on the sample holder of FIG. S-polarized light is applied to the remaining surface of the thermal oxide film (upper half surface) and the surface of the film removed (lower half surface) at an oblique incidence (incident angle of 70 degrees). Measure and call them T (SiO ₂ / Si) and T (Si), respectively. Here, the ratio of the transmission spectrum of the thin film sample on the substrate to the substrate (relative transmittance = T (SiO ₂ / Si) / T (Si)) was obtained. The result is shown in FIG. The “curve in which the maximum value and the minimum value are adjacent” appearing in this figure agrees qualitatively well with the calculation results of FIGS. 5 (a) and 5 (b).

  FIG. 7 shows the measurement results of the same sample as in paragraph [0035] with the incident angle changed. In this measurement result, there is almost no difference in the relative transmittance when the incident angle is between 0 ° and 40 °, and it appears in the relative transmittance when the incident angle is increased to 60 °, 70 °, 80 °, and 85 °. The structure is growing. This incident angle dependence is also in good agreement with the calculation results of FIGS. 5 (a) and 5 (b).

  A silicon substrate having a uniform thickness of 700 μm was set on the sample holder shown in FIG. 1, and the incident angle on the sample was set to 85 degrees. Next, while rotating the sample around an axis perpendicular to the sample surface and passing through the center, the transmission spectrum of the upper half surface and the lower half surface of the sample are measured at each angle, and the ratio of the spectra (relative) Transmittance) was calculated. The result is shown in FIG. In this figure, the rotation angle around the central axis of the sample is used as a parameter. The “curve in which the maximum value and the minimum value are adjacent” shown in this figure is because the substrate is not completely flat.

  The present invention is a technique for measuring a complex dielectric constant of a conventional material or a new material in a high-frequency region (30 GHz to 3 THz) that is expected to be developed in the future. In this area, it is indispensable to measure the complex dielectric constant of a thin film in order to develop a new high-frequency device, and the present invention can measure the complex dielectric constant even if the thickness of the thin film is 1 μm or less in this frequency area. Since it is a technology, it can be put to practical use as an apparatus for in-situ product management at the actual manufacturing site of the semiconductor industry. Furthermore, since the flatness of the semiconductor wafer actually used in the semiconductor industry is larger than 1 μm, it becomes a simple high-sensitivity complex dielectric constant measuring device as in the prior art. Provided is a method and an apparatus capable of measuring a complex dielectric constant of a thin film with the same measuring apparatus and measuring the complex dielectric constant even when the thickness of the thin film is 1 μm or less.

Conceptual diagram of complex permittivity measurement device Diagram showing transmittance spectrum of substrate and thin film on substrate at normal incidence Diagram showing the incident angle dependence of the transmittance and reflectance of S and P deflections Diagram showing the frequency dependence of the transmittance spectrum of S deflection when the incident angle is changed Diagram showing relative transmittance when film thickness is 2μm and 0.4μm Relative transmittance measurement result at 70 ° incidence with S deflection Measurement result of relative transmittance of S-polarized light at different incident angles Measurement results of relative transmittance when the measurement location is changed on the sample surface at an incident angle of 85 degrees

Explanation of symbols

10 Complex permittivity measurement system
11
sample
12
light source
13
Detector
14
Mechanical chopper
15,17,19,21 lens
16,18,20,22 Aperture
30 Incident system
31 Light receiving system

Claims (10)

  In the dielectric constant measuring apparatus, the sample is irradiated with light, the transmitted light transmitted through the sample is measured, and the flatness and complex dielectric constant of the sample are obtained based on the spectrum of the transmitted light. Dielectric constant measuring device.   2. The dielectric constant measuring apparatus according to claim 1, wherein a position of incident light on the sample is variable, and a position of a detector that receives the transmitted light is also variable.   3. The dielectric constant measuring apparatus according to claim 2, wherein an incident angle of incident light on the sample is variable.   In the dielectric constant measurement method, the sample is irradiated with light, the transmitted light transmitted through the sample is measured, and the flatness and complex dielectric constant of the sample are obtained based on the spectrum of the transmitted light. Dielectric constant measurement method.   5. The dielectric constant measurement method according to claim 4, wherein a spectrum of the transmitted light is measured while changing a position of incident light on a sample having a uniform dielectric constant on the sample surface, and the spectrum at a reference position is measured. A dielectric constant measuring method, wherein the flatness of the sample is obtained by comparison.   6. The dielectric constant measuring method according to claim 5, wherein the flatness of the sample is obtained by increasing an incident angle of the incident light to the sample.   7. The dielectric constant measurement method according to claim 4, wherein a complex dielectric constant of the thin film is obtained by using a substrate having a uniform dielectric constant and a uniform thickness and a sample in which a thin film is provided on a part of the substrate. A dielectric constant measurement method characterized by:   8. The dielectric constant measurement method according to claim 7, wherein the complex dielectric constant of the thin film is obtained by irradiating the substrate portion and the thin film portion of the sample with light and comparing the spectra of the light transmitted through the substrate portion and the thin film portion. A dielectric constant measurement method.   9. The dielectric constant measuring method according to claim 4, wherein the irradiated light is S-polarized light. 10. The dielectric constant measurement method according to claim 4, wherein the wavelength of the irradiated light is millimeter wave, submillimeter wave, or terahertz light.

Images