JP3366525B2 - Film thickness measuring method and film manufacturing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for measuring a film thickness by measuring an X-ray reflectivity from a film and a method for manufacturing a film using the method.

[0002]

2. Description of the Related Art In recent technical fields such as various thin films such as an oxide film, a metal film for an electrode, a metal film for a wiring, and a dielectric film in a large-scale semiconductor device (LSI), and a metal thin film in a magnetic head. As the number of layers increases, the required accuracy of film thickness control is becoming more and more strict.

For example, in the case of a barrier metal in an LSI, a silicon oxide film, a titanium film, and a titanium nitride film are sequentially stacked on a silicon substrate. 10 nm or less to several tens
For a titanium film and a titanium nitride film having a thickness of nm, control with an accuracy of several angstroms is required. As a measurement technique required for such film thickness control, it is required that the measurement can be performed with high accuracy, simple, short time, and that the measured film thickness is uniquely determined. Further, when a plurality of films are successively laminated by the same film forming apparatus such as a barrier metal in an LSI or a metal film in a magnetic head, it is necessary that each film thickness can be measured in a multilayer film state.

On the other hand, it has been clarified that the film density and surface irregularities of a gate oxide film and an interlayer insulating film having a low dielectric constant in an LSI greatly affect device characteristics. As described above, in addition to the film thickness, the film density and the surface unevenness can be used as indices for controlling the film quality. Therefore, in order to manage the manufacturing process of forming a film, there is a need for the development of a measurement technique for quickly measuring physical quantities such as film thickness, film density, and surface irregularities with high accuracy.

[0005] A conventionally known film thickness measuring technique will be described generally. Most commonly, an optical ellipsometer is used for film thickness measurement. Polarized light is incident on the object to be measured, and the thickness and refractive index of the thin film are measured from the change in the polarization state of the reflected light. However, the method using an ellipsometer can be applied only to a material through which light passes, such as a silicon oxide film, and cannot be applied to measurement of an opaque film such as a barrier metal in an LSI described above or a metal film in a magnetic head. .

In the case of a silicon oxide film, 64M
For extremely thin films with a thickness of 10 nm or less, such as gate oxide films used in DRAMs, the measurement accuracy rapidly deteriorates.
Not available. The ellipsometer calculates the film thickness on the assumption that the film density is constant. However, in practice, even with the same material, the film density generally differs depending on the film forming method, film forming conditions, and film thickness. For example, in the case of a thin silicon oxide film, the film density greatly differs depending on the film forming method and the film forming conditions. For this reason, the ellipsometer cannot measure the film thickness accurately.

[0007] Other conventional methods for measuring film thickness include fluorescence analysis, RBS (Rutherford Backscattering Spectrometry),
AES method (Auger electron spectroscopy: Auger Electron Spect
roscopy), XPS method (X-ray photoelectron spectroscopy: X-ray Photo
electron spectroscopy), SIMS (Secondary Ion Mass Spectrometry), X-ray reflectivity, and the like.

Among these measurement methods, the fluorescence analysis method measures the fluorescent X-ray emitted when a sample is irradiated with X-rays, determines the amount of the contained element from the intensity of the fluorescent X-ray, and determines the film thickness. It is a conversion method. In this fluorescence analysis method, only the amount of the element is measured, so that the obtained film thickness is a converted film thickness, and is not necessarily the actual film thickness. Also, a laminated film containing the same element, for example, Ti (titanium) and TiN (titanium nitride)
It is not suitable for measurement of laminated films and the like.

The RBS method is a method of irradiating a sample with high-energy light ions and analyzing the energy distribution of the back-scattered light ions to measure the film thickness. In the RBS method, it is difficult to separate the energy of ions that are backscattered by target atoms having close atomic mass, and thus it is difficult to evaluate a stacked film made of elements having close atomic mass. In addition, the device itself is expensive.

In the measurement of the film thickness by the AES method or the XPS method, the composition of a substance in the depth direction of the sample is examined by performing AES or XPS analysis while sputtering the sample surface with ions such as Ar to obtain a film. It is a method of converting to thickness. The measurement of the film thickness by the SIMS method is as follows.
In this method, the composition of a substance in the depth direction of a sample is examined by performing mass analysis of secondary ions generated by ion irradiation, and the result is converted into a film thickness.

Since the film thickness measuring method using the AES method, the XPS method, and the SIMS method is a so-called destructive inspection in which the sample surface is analyzed while being sputtered, the product itself cannot be inspected. There is complexity in the process of sputtering. In contrast to these film thickness measurement methods, the X-ray reflectivity method of irradiating a sample with X-rays and measuring the film thickness by X-ray interference on the sample is non-destructive and has a high accuracy of about 0.1 nm resolution. Since this method can evaluate the thickness of a metal, a silicon oxide film, a silicon nitride film, or the like, it is a desirable method as a thin film thickness measuring means. In particular, in the Fourier method of evaluating the film thickness by performing the Fourier transform on the interference fringe appearing in the X-ray reflectivity, the film thickness can be measured quickly and uniquely.

Next, a conventional method for measuring the film thickness by the X-ray reflectivity method using the Fourier method will be described. First, the sample is irradiated with X-rays while changing the incident angle, and the reflectance is measured. Thereby, a curve R having an interference fringe is obtained. Next, in order to extract only the interference vibration of the curve R,
An average reflectance curve R _ave is obtained from the curve R, and the average reflectance curve R _ave is subtracted from the curve R. Thus, a vibration component curve Δ (θ) having only interference vibration is obtained.

Subsequently, the vibration component curve Δ (θ) is Fourier-transformed to obtain a relationship F (d) between the Fourier coefficient F and the film thickness d. Since the peak position of the curve thus obtained corresponds to the thickness of the film formed on the sample surface, the film thickness can be uniquely determined by measuring the peak position of the curve. In order to obtain a more accurate film thickness in the above series of film thickness measurement procedures, how to obtain an average reflectance curve R _ave
It is important to decide. Conventionally, (1) an average averaging operation of the data of the curve R is repeated to obtain an average reflectance curve _Rave . (2) An arbitrary range is designated in the curve R, and an average reflectance curve _Rave is obtained by polynomial approximation. (3) Give some points near the average from the data of the curve R, and obtain an average reflectance curve R _ave by interpolating the values with a spline curve. (4) A single element consisting only of the constituent elements of the thin film or the substrate A method of finding an average reflectance curve R _ave by fitting a theoretical formula of reflectance in a layer sample to measured data, or the like, has been used.

However, in the above-mentioned conventional method for measuring the film thickness by the X-ray reflectivity method, since the X-ray reflectivity curve R itself has a feature of sharply decreasing near the critical angle of total reflection, an appropriate average reflectivity curve R Seeking _ave was not easy. That is, in the method (1), the characteristics themselves may be averaged by averaging the curve R. In addition, since the number of times of averaging is arbitrary, there is a possibility that the arbitraryness is reflected in the measured film thickness itself.

In the above methods (2) and (3), the conditions are arbitrarily selected, such as designating a range, and there is a possibility that interference vibrations may be absorbed. Further, in the above method (4), the function type of the reflectance is correct, but it is necessary to provide the film density, the refractive index, and the like in obtaining the reflectance, and the handling is complicated and versatile. Further, when measuring the film thickness of the laminated film, there is arbitrariness in which film element is used as a reference.

On the other hand, as a conventional method for measuring the density of a film, there has been only a classical method of calculating from a weight measurement. That is, the weight and thickness of the film to be measured are measured after the film is formed, and then only the film is chemically removed by etching, and the weight is measured again. The weight of the film itself is calculated from the weight difference before and after the etching removal. Separately, the volume is calculated from the film thickness and the area, and the film density is calculated from the weight of the film itself.

In such a conventional method for measuring the film density,
The film density is determined by the film-forming process, because it is limited to materials that only chemically remove the film by etching, the measurement accuracy is extremely poor for thin films, and the measurement is too time-consuming and impractical. It was not possible to use it for management in.

[0018]

As described in detail above,
In the prior art, a film thickness measurement technique suitable for the demanded precision of recent strict film thickness control has not been established, and it is desired to establish a film thickness measurement technique that can be used in a film manufacturing method. In addition to film thickness, film density and surface irregularities are also important as indices for controlling film quality. Techniques for accurately and non-destructively measuring these film densities and surface irregularities are also expected.

An object of the present invention is to provide a method for measuring the thickness of a film, which can be measured with high accuracy, is simple, can be performed in a short time, and can uniquely determine a measured film thickness. Another object of the present invention is to provide a method for measuring the thickness of each layer in the state of a multilayer film. Another object of the present invention is to provide a film manufacturing method capable of adjusting the film forming conditions using the measurement result obtained by the film thickness measuring method.

[0020]

SUMMARY OF THE INVENTION The object of the present invention is to provide a method for producing X
In a film thickness measuring method of measuring a line reflectivity, extracting interference vibration from the measured X-ray reflectivity, and calculating a film thickness of the film by performing a Fourier transform of the interference vibration, the surface of the film has irregularities. In addition, an analytical expression obtained by adding a constant term representing the background to a term consisting of a product of a power function of the incident angle representing the attenuation of the reflectance in the absence of the incident angle and an exponential function representing the influence of the surface unevenness of the film, This is achieved by a film thickness measuring method, wherein an average reflectance is obtained by fitting an X-ray reflectance, and the interference vibration is extracted using the measured X-ray reflectance and the average reflectance. By doing so, the analytic formula incorporates the property of reflectance,
Conventionally, the degree of freedom that has been arbitrarily set can be limited in a correct manner. Further, even when the structures of the films are different, the analytic formula can be used with versatility and the measurement can be performed with high accuracy. In addition, the speed of measurement can be improved.

In the above-mentioned film thickness measuring method, the analytical expression is that the intensity is I ₀ , the incident angle of the X-ray is θ, the origin correction value of the incident angle is θ ₀ , and the root mean square value of the roughness of the film surface is Assuming that σ, the wavelength of the X-rays are λ, and the background constant is B ₀ , I ₀ (θ−θ ₀ ) ^-4 · expｅｘ− (4πσ / λ · sin
θ) ² ｝ + B ₀ .

In the above-described film thickness measuring method, it is preferable to estimate the roughness of the film surface from the root mean square value of the roughness obtained by fitting the measured X-ray reflectance to the analytical expression. This allows the analysis
The roughness of the film surface can be estimated from the root mean square value of the roughness obtained by fitting the measured X-ray reflectivity to the equation . In the above-described film thickness measuring method, it is desirable that the interference vibration is obtained by dividing the measured X-ray reflectance by the average reflectance.

In the film thickness measuring method described above, when the film is a multilayer film having two or more layers, the incident angle is low with respect to a layer located on the surface side of the film, For the layer located at the position (1), the range of the incident angle may be limited based on the position of the layer to be measured in the multilayer film, so that the measurement is performed. Thus, the range of the angle measurement region can be limited according to the purpose of the measurement, and the desired depth can be analyzed.

The object of the present invention is to provide a method for manufacturing a film for forming a film on an underlying substrate, wherein the film thickness of the film is measured by the above-described film thickness measuring method, and the film thickness is measured by using the measured film thickness. This is achieved by a method for producing a film, characterized by adjusting the film forming conditions. The object of the present invention is to provide a method for manufacturing a film for forming a film on an underlying substrate, wherein the film thickness of the film is measured by the above-described film thickness measuring method, and the measured film thickness is used as an initial value to determine the X-ray reflectivity. By calculating the calculated X-ray reflectivity and the measured X-ray reflectivity, the film density of the film, at least one physical quantity of the surface unevenness is obtained, and the film thickness, the film density, and the surface of the film are obtained. This is achieved by a method for producing a film, characterized in that the film-forming conditions of the film are adjusted using at least one physical quantity of unevenness.

In the above-mentioned method for producing a film, an oxidation treatment, a nitridation treatment, an oxynitridation treatment, a sputtering treatment, a CVD treatment, an electron beam Deposition processing,
In a manufacturing apparatus that performs any one of ion beam evaporation processing, molecular beam evaporation processing, and liquid phase growth processing, a film forming temperature,
At least one parameter of the gas introduction amount and the gas mixture ratio may be adjusted.

The object of the present invention is to provide a method of etching a film for removing at least a part of a film formed on a base substrate, by measuring the film thickness of the film etched by the above-mentioned film thickness measuring method. The etching condition of the film is adjusted by using the film thickness of the film.

[0027]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A method for measuring a film thickness according to a first embodiment of the present invention will be described with reference to FIGS. FIG. 1 is a view for explaining a film thickness measuring method according to the present embodiment, FIGS. 2, 4 to 6 are graphs showing results measured by the film thickness measuring method according to the present embodiment, and FIG. 3 is a conventional film thickness measuring method. 6 is a graph showing the results measured by the method.

First, an outline of the film thickness measuring method according to the present embodiment will be explained with reference to FIG. First, the surface of the sample on which the film to be measured is formed is irradiated with X-rays to measure the X-ray reflectivity , thereby obtaining a curve R (step S1). Next, an average reflectance curve R _ave is determined based on the curve R measured in step S1 (step S2).

Subsequently, an average reflectance curve R _{ave is obtained} from the curve R.
Is subtracted to extract a vibration component curve Δ (θ) reflecting the interference vibration of the curve R (step S3). Thereafter, the vibration component curve Δ (θ) is subjected to variable conversion to obtain a vibration component curve Δ (x) (step S4). Next, the vibration component curve Δ (x) is Fourier-transformed (step S5),
The film thickness is measured from the peak position of the obtained Fourier coefficient F (d) (step S6).

In the film thickness measuring method using the X-ray reflectivity method performed as described above, in the present embodiment, when the average reflectance curve R _ave is obtained in the above film thickness measuring method,
Using an analytical expression consisting of a term consisting of the product of a power function of the incident angle representing the attenuation of the reflectance when there is no unevenness on the sample surface and an exponential function representing the influence of the surface unevenness, and a constant term representing the background It is characterized by performing data fitting.

The inventor of the present application has newly found that the average reflectance curve R _ave can be fitted accurately and without giving any arbitrariness by using such an analytical expression. Specifically, for example, an analytical expression represented by R = I ₀ (θ−θ ₀ ) ⁻⁴ · exp ｛− (4πσ / λ · sin θ) ² } + B ₀ (1) can be used.

Where I ₀ is the intensity, θ is the incident angle of the X-ray,
θ ₀ is the origin correction value of θ, σ is the root mean square (rms) roughness of the sample surface (hereinafter referred to as rms (root mean square) roughness), λ is the wavelength of X-rays, and B ₀ is the background constant. . The background constant B ₀ includes zero. The angle of incidence referred to in the specification of the present application is not an angle with respect to a normal to the surface of the sample, but an angle formed between the surface of the sample and the incident X-ray.

In this embodiment, as a power function of the incident angle representing the attenuation of the reflectance when the sample surface has no irregularities, I ₀ (θ
The function of −θ ₀ ) ⁻⁴ is used because the reflectance of a sample having no irregularities on the surface has a property of decreasing at θ ⁻⁴ with respect to the incident angle θ of the X-ray to the sample. . Exponential function representing the effect of surface irregularities is exp exp
The function of {− (4πσ / λ · sin θ) ² } is used. Assuming that the rms roughness is σ and the Debye waller type roughness is assumed, the attenuation of the reflectance due to this roughness is approximately exp ｛− (4π
σ / λ · sin θ) ² }.

Therefore, by using the term composed of these products, it is possible to construct an analytical expression incorporating the property of the reflectance, and it is possible to restrict the degree of freedom, which has been conventionally set arbitrarily, in a correct manner. it can. This analytic expression is given by I
_Since it is a monotonically decreasing function having four parameters of ₀ , θ ₀ , σ, and B ₀ , there is no possibility of absorbing the vibration of the curve R during fitting.

Further, since these parameters do not include constants peculiar to the sample, they have versatility to be used for various samples. Further, when the curve R is fitted by this equation, information on the sample surface roughness can be obtained simultaneously from the obtained value of σ. Note that the curve fitting may be performed using, for example, the least squares method. Silicon oxide film: SiO ₂ Next, data obtained by simulating the X-ray reflectivity will be described by comparing results obtained by analyzing the data obtained by the conventional method and the film thickness measurement method according to the present embodiment.

FIG. 2 is a graph showing the result analyzed by the film thickness measuring method according to the present embodiment, and FIG. 3 is a graph showing the result analyzed by the conventional method. For analysis, the film thickness was 5 nm,
The result of simulating the X-ray reflectivity of an ultra-thin silicon oxide film having a surface roughness of 0.4 nm was used. From the reflectance obtained by the simulation, FIG.
The curve R shown in FIG. 3A was obtained, and the thickness of the silicon oxide film was obtained by the conventional method and the film thickness measuring method according to the present embodiment.

In the film thickness measuring method according to the present embodiment shown in FIG. 2, the average reflectance curve R _ave is fitted by the least squares method using the analytical formula (1), and in the conventional method shown in FIG. The average reflectance curve R _ave was obtained by smoothing the curve R for each individual region (step S2).
The value of each parameter obtained at the time of fitting of the analytical expression (1) is as follows: I ₀ = 7.98 (8) θ ₀ = 0.001 deg (0 deg) σ = 0.397 nm (0.4 nm) B ₀ = 5.16 (5) The numerical values in the parentheses are parameters actually given in the simulation. In the film thickness measurement method according to the present embodiment, the rms roughness σ can be obtained at the time of fitting the average reflectance curve R _ave , but the value is 0.397 nm, and a correct value can be obtained within the analysis error. did it.

Next, the curve R and the average reflectance curve R _ave
Thus, the vibration component of the curve R was extracted (step S3).
Oscillation component curve delta (theta) was calculated using the Δ (θ) = logR (θ ) -logR ave (θ) ... (2). Subsequently, the horizontal axis variable conversion of the vibration component curve Δ (θ) was performed (step S4). Variable conversion was performed using x = √ (θ ² −θ _c ² ) / λ (3). Here, λ is the wavelength of the X-ray, and θ _c
Is the critical angle for total reflection.

Thereafter, the vibration component curve Δ obtained by performing variable conversion
(X) was subjected to Fourier transform to obtain Fourier coefficients F (d) shown in FIGS. 2B and 3B (step S5). As a result of measuring the film thickness of the silicon oxide film from the peak position of the graph obtained as described above, when the film thickness measuring method according to the present embodiment was used, a film thickness of 5.1 nm was obtained. In the conventional method, the deviation was as large as 6.1 nm. Multilayer film: TiN / Ti / SiO ₂ / Si Next, a silicon oxide film having a thickness of 40.8 nm, a Ti film having a thickness of 7.7 nm, and a film thickness of 15.6 n are formed on a silicon substrate.
FIGS. 4 to 6 show the results obtained by analyzing the curve R obtained by the experiment using the sample in which the mN TiN films are sequentially laminated by the film thickness measuring method according to the present embodiment.

FIG. 4 shows the results of the analysis in the entire angle range, FIG. 5 shows the results of the analysis in the angle range of 0 to 2 °, and FIG. 6 shows the results of the analysis in the range of 2 to 6 °. is there. In the analysis, the average reflectance curve R _ave was obtained by fitting the curve R to the analysis formula (1) using the least squares method. The value of each parameter obtained by the analysis was I ₀ = 8.34 θ ₀ = 0.002 deg σ = 0.287 nm B ₀ = 11.06. As a result, the rms roughness of the sample surface is set to 0.
It was found to be 387 nm.

As shown in FIGS. 4 (a), 5 (a) and 6 (a), a smooth average reflectance curve R _ave can be obtained regardless of the angle range of the fitting. It turns out that there is. Based on the average reflectance curve R _ave thus obtained, a vibration component curve Δ (x) was obtained using Expressions (2) and (3), and the result of Fourier transform was obtained. 5 (b) and 6 (b) were obtained.

In the graph shown in FIG. 4B, an interference peak indicating a Ti film, an interference peak indicating a Ti film and a TiN film, and an interference peak indicating a silicon oxide film are observed. These peak positions are 7.5 nm and 23.8, respectively.
nm and 40.6 nm. Accordingly, it was determined that the thickness of the Ti film was 7.5 nm, the thickness of the TiN film was 16.3 nm, and the thickness of the silicon oxide film was 40.6 nm. As described above, it was found that by using the film thickness measuring method according to the present embodiment, detection was possible with a measurement accuracy of about 0.5 nm in error.

FIG. 5 showing an analysis in the low angle region.
In the result of (b), the interference peak obtained from the Ti film alone and the interference peak obtained from the silicon oxide film disappear, and only the interference peak indicating the TiN film and the interference peak indicating the Ti film and the TiN film are reduced. Has been detected. On the other hand, in the result of FIG.
The interference peak obtained from the N film disappears, and only the interference peak obtained from the Ti film and the interference peak obtained from the silicon oxide film are detected.

The reason why the peak detected by changing the analysis angle region changes as described above is that the X-rays are incident deeper into the sample as the incident angle is larger, and the information deeper into the sample as the incident angle is larger. Is obtained. Therefore, if the area is limited to the range of the angle measurement area according to the purpose of the measurement, the desired depth can be analyzed. For example, when measuring the upper layer film, the analysis range can be limited to a low angle region, and when measuring the lower layer film, the analysis range can be limited to a high angle region. Gate oxide film: SiO ₂ Next, the result of analyzing the gate oxide film formed on the silicon substrate under the following film forming conditions by the film thickness measuring method according to the present embodiment will be described.

The film formation conditions were as follows: a film formation temperature of 1000 ° C., an atmosphere gas of oxygen (200 Torr), and a thermal oxidation time of 30 minutes.
In about a minute, a gate oxide film having a thickness of about 4 nm was formed. In measuring the X-ray reflectivity, X-rays having a wavelength of 1.3 angstrom were used, and the X-ray reflectivity from the gate oxide film was measured in a range of a scattering angle (twice the X-ray incident angle) of 0 to 7 degrees. It was measured.

FIG. 7 shows the result of analysis by the film thickness measuring method according to the present embodiment. In the analysis, the average reflectance curve R _ave was obtained by fitting the curve R to the analysis formula (1) using the least squares method. The value of each parameter obtained by the analysis was I ₀ = 6.93 θ ₀ = 0.002 deg σ = 0.37 nm B ₀ = 10.5. As a result, the rms roughness of the sample surface is set to 0.
It was found to be 37 nm.

As shown in FIG. 7A, a smooth average reflectance curve R _ave was obtained. Based on the average reflectance curve R _ave obtained in this manner, the vibration component curve Δ (x) was obtained using the equations (2) and (3), and the result of Fourier transform was obtained. The graph of the Fourier coefficient F (d) shown in FIG. In the graph shown in FIG. 7B, a clear interference peak indicating the gate oxide film was observed, and the peak position was 3.9 nm. Therefore, the measurement result that the thickness of the gate oxide film was 3.9 nm was obtained.

The thickness of the gate oxide film in the DRAM is about 10 nm in a 64M-DRAM, and 1 G-DRA
M is expected to be about 6 to 7 nm. Conventionally, there has been no method capable of accurately and easily measuring the thickness of the gate oxide film. In the conventional method using an ellipsometer, accuracy is rapidly deteriorated at a film thickness of 10 nm or less. On the other hand, according to the present embodiment, even for an extremely thin gate oxide film having a thickness of 10 nm or less, accurate film thickness measurement can be performed simply, in a short time. Further, the film formation process can be controlled based on the measured film thickness. Titanium Nitride Film: TiN (Film Thickness) Next, the result of analyzing the titanium nitride film formed on the silicon substrate under the following film forming conditions by the film thickness measuring method according to the present embodiment will be described.

The conditions for forming the titanium nitride film by sputtering are as follows.
W, atmosphere gas is nitrogen, gas flow rate is 10 sccm (Stan
dardCubic Centimeter per Minute). In measuring the X-ray reflectivity, X-rays having a wavelength of 1.3 angstroms were used, and the X-ray reflectivity from the titanium nitride film was measured at an X-ray incident angle of 0 to 6 degrees.

FIG. 8 shows the result of analysis by the film thickness measuring method according to the present embodiment. Similar to the procedure described above,
The measured value of the X-ray reflectivity was analyzed to obtain a clear vibration component curve Δ (x) as shown in FIG. As a result of Fourier transform of the vibration component curve Δ (x), a graph of Fourier coefficient F (d) shown in FIG. 8B was obtained. FIG.
In the graph shown in (b), a very clear interference peak indicating the titanium nitride film was observed, and the peak position was 21.4 n.
m. Therefore, the thickness of the titanium nitride film is 21.
A measurement result of 4 nm was obtained.

The time required for the fitting analysis according to the present embodiment was extremely short, about several minutes. Since the titanium nitride film is an opaque metal film, it is impossible to measure the film thickness using an ellipsometer. Further, since it is a compound, it is also difficult to calculate the film thickness from the fluorescent X-ray intensity. Although it is possible to measure the film thickness using an electron microscope, it is very troublesome, so that it is effective for acquiring basic data, but cannot be used as a measurement technique for controlling film formation.

Moreover, in the case of a film forming method by sputtering such as a titanium nitride film, the state of film formation is slightly changed depending on film forming conditions such as an RF voltage, a gas flow rate, and a gas component ratio. Therefore, when performing a series of film formation under the same conditions,
Fine control of film formation is required. According to this embodiment, it is possible to easily and accurately measure the film thickness in a short time. For example, a film is formed using an inspection substrate at the beginning of a series of film forming steps, and the film thickness is measured. Fine adjustment of film formation conditions is possible from the values. Further, by monitoring the measurement data over a long period of time, it is possible to cope with a change with time of the film forming apparatus. Multilayer film: TiN / Ti / SiO ₂ / Si Next, a sample obtained by sequentially laminating a silicon oxide film, a Ti film, and a TiN film on a silicon substrate was analyzed by the film thickness measuring method according to the present embodiment. Will be described.

FIG. 9 also shows the result of analysis by dividing the angle region into three. FIG.
As a result of performing analysis in an angle region of 2 °, FIG.
FIG. 12 shows the result of the analysis in the angle range of 4 to 6 ° as a result of the analysis in the angle range of ４4 °.
In the analysis, the average reflectance curve R _ave was obtained by fitting the curve R to the analysis formula (1) using the least squares method.

The value of each parameter obtained by the analysis was I ₀ = 7.2 θ ₀ = −0.002 deg σ = 1.6 nm B ₀ = 15.3. Thereby, the rms roughness of the sample surface is 1.
It was found to be 6 nm.

FIGS. 9A, 10A and 11
12 (a) and FIG. 12 (a), it can be seen that a smooth average reflectance curve R _ave is obtained when fitting is performed in any angle range. Based on the average reflectance curve R _ave obtained in this manner, a vibration component curve Δ (x) was obtained using the equations (2) and (3). Features in each angle range will be described.

As shown in FIG. 10A, in a relatively low angle range of 0.6 to 2 °, the Ti film modulated by the Ti film is used.
Interference vibration of the N film was observed. As shown in FIG. 11A, in the middle angle range of 2 to 4 °, interference vibration of the Ti film in which short-period modulation by the SiO ₂ film overlaps was observed. As shown in FIG. 12A, in the relatively high angle range of 4 to 6 °, monotonous interference vibration of the lowermost SiO ₂ film was observed.

From the analysis results of the entire angle range shown in FIG. 9, it is found that there are differences in the characteristics of the interference vibration in each of these angle ranges. As a result of performing the Fourier transform on the interference vibration shown in FIGS. 10A, 11A, and 12A, FIG.
(B), graphs of Fourier coefficients F (d) shown in FIGS. 11 (b) and 12 (b) were obtained.

As shown in FIG. 10B, 0.6 to 2 °
At a relatively low angle range of 135.2 nm, T
A peak corresponding to the sum of the thicknesses of the i film and the TiN film is clearly seen.
Was done. In this multilayer film, the density between the Ti film and the TiN film
Compared to the difference, Ti film and SiO _TwoLarge density difference between membranes
Therefore, the sum of the thicknesses of the Ti film and the TiN film becomes a clear peak.
You. As shown in FIG. 11B, the middle corner of 2 to 4 °
In the temperature range, a pin corresponding to a Ti film is located at a position of 20.6 nm.
Is clearly seen, and SiO at 99.7 nm_Twofilm
Was clearly seen. Equivalent to TiN film
No peak appears. As shown in FIG.
In the relatively high angle range of 4 to 6 °, the order of 99.8 nm
SiO_TwoOnly the peak corresponding to the film is clearly seen
Was.

From these results, the thickness of the TiN film was 11
The analysis result was 4.6 nm, the thickness of the Ti film was 20.6 nm, and the thickness of the SiO ₂ film was 99.8 nm. These values are the exact film thicknesses obtained by performing fitting by trial and error assuming all parameters such as film thickness, film density, irregularities, etc. (TiN film thickness is 115.1 nm, Ti film is Has a thickness of 20.4 nm and a SiO ₂ film has a thickness of 99.9 n.
Compared with m), the agreement was within ± 1%.

For comparison, fitting was performed for all angle regions without dividing the regions. As a result, a vibration component curve Δ (x) as shown in FIG. 13 was obtained. As a result of Fourier transform of this vibration component curve Δ (x), Ti
The thickness of the N film is 117.1 nm, and the thickness of the Ti film is 20.4.
nm, the film thickness of the SiO ₂ film was 100.3 nm, which was significantly different from the accurate film thickness.

As described above, in the case of a multilayer film having two or more layers, if the structure of the multilayer film (the order of lamination) and the material and thickness of each film are determined, the interference vibration of which layer is applied to which incident angle region. You can decide whether it will be reflected. Therefore, if the target value of the measured film thickness of the multilayer film is known, it is possible to predict which angle region should be analyzed for which film from the structure and material of the multilayer film.

Further, by dividing the incident angle into regions, in the analysis of the X-ray reflectivity, the mixing of extra vibration components, which become noise, is reduced. Therefore, the peak in the Fourier intensity curve after the Fourier transform is obtained. And the film thickness can be easily identified. Usually, in each film forming process of the multilayer film, naturally, the structure of the multilayer film, the material of each film, and the film thickness are set as target values. Therefore, by storing the optimal angle region for the multilayer film as a database and using the stored data before actually analyzing the measured values and calculating the film thickness, it is possible to calculate the film thickness more quickly and accurately. It is possible. In measuring the X-ray reflectivity, the angle range of the incident angle and the angular accuracy in accordance with the film thickness according to the purpose are also obtained.
By holding as a database, more efficient and quick measurement is possible.

As a conventional method for measuring the thickness of a metal thin film, a fluorescent
There is a method of calculating from the X-ray intensity using a standard sample,
A multilayer film containing a common element called Ti (TiN / Ti / S
iO _Two/ Si) was not possible. In contrast,
According to this embodiment, a common element is included as a material of each film.
Accurate film thickness measurement is also possible for a multilayer film. This
This controls the continuous film forming process for manufacturing multilayer films,
The present embodiment enables management for the first time.

The stacked structure of TiN / Ti is used as a silicide electrode or a protective film for Al wiring. The silicide electrode is for thermally reacting with Si at an electrode opening to produce a titanium silicide low-resistance electrode. The largest influence on the electrode resistance is the thickness of the Ti film. T
If the film thickness of the i film is too thick, the electrode pattern spreads, and if the film thickness of the Ti film is too thin, the electrode resistance increases.

In the protective film for the Al wiring, the thickness of the Ti film affects the reliability and electric resistance of the wiring. Therefore, when a Ti film is formed by sputtering, it is necessary to suppress the error of the film thickness to within several angstroms by comparing the electric characteristics of the film with the observation result of the Ti film by an electron microscope. Conventionally, there is no film thickness measurement technology for directly managing the film forming process, and at present, the cause is analyzed retroactively based on the result of characteristic failure or the like after manufacturing a semiconductor element. According to the present embodiment, the film thickness of such a multilayer structure can be easily and accurately measured. Therefore, at the beginning of a series of film-forming processes, a film is formed using an inspection sample, and the film thickness of the formed multilayer structure is measured, so that the measurement result can be reflected in the film-forming conditions. In addition, the management technology of the film forming process is remarkably improved.

Therefore, by using the film thickness measuring method according to the present embodiment, it is possible to efficiently form a silicide electrode or an Al wiring having a desired resistance value. Titanium nitride film: TiN (density) The measurement of the film density of the titanium nitride film according to the present embodiment will be described. Although the measurement of the thickness of the titanium nitride film has been described above, the denseness of the titanium nitride film used as the protective film is an important point of the film quality. For example, when a titanium nitride film is used as a protective film for a multilayer wiring, WF ₆ used when filling the connection hole between the multilayer wirings by CVD growth of tungsten penetrates the titanium nitride film as a protective film, and the titanium Peeling may occur between the film and the aluminum film. Although the film density is the most appropriate as an index indicating the denseness of the film, it has not been possible to measure the density of a thin film of about several tens nm used as a protective film for wiring.

Using the accurate film thickness measured by the film thickness measuring method of this embodiment as an initial value, the film thickness, the film density, and the surface unevenness can be calculated by a fitting method using the film thickness, the film density, and the surface unevenness as parameters. In order to perform accurate and rapid analysis by the fitting method, it is essential to select an appropriate initial value. Conventionally, a design value was input as an initial value. However, in the present embodiment, an accurate measured film thickness can be used as an initial value, so that accurate and short-time processing can be performed. Furthermore, in the conventional fitting technique, since the reflection curve is sharply attenuated, there is a problem that it is difficult to grasp the quality of the fitting result. However, according to the present embodiment, since fitting can be performed only with the vibration component calculated first,
The quality of the fitting result can be clearly understood.

A general fitting method is described in the following paper, for example. (1) L. G. FIG. Paratt, Phys. Rev .. ,
95, 359 (1954) (2) L.P. Nevot and P.S. Crosse, R
ev. Phys. App1. , 15, 761 (198
0) (3) B. Vida1 and P.M. Vincent,
App 1.0 ptics, 23, 1794 (198
4) (4) S.P. K. Sinha, E .; B. Shirota
and S.M. Garoff, Phys. Rev ..
B, 38, 2297 (1988). Chen and M.S. Heald, J.M.
App1. Phys. , 66 (4), 15,1793
(1989) Therefore, according to the present embodiment, not only the film thickness but also the film density greatly affects the characteristics of the titanium nitride film.
For the first time, it is possible to control the film forming process at the practical level by the film thickness and the film density. SiOF Film A specific example in which the present embodiment is used for controlling the film formation of a SiOF film will be described.

The low dielectric SiOF film developed for the purpose of preventing the wiring delay due to the parasitic capacitance is formed by further adding fluorine (F) when forming the SiO ₂ film. Generally, Si
The OF film has a problem that the hygroscopicity is increased by the addition of fluorine and the reliability is lacking. However, by supplying sufficient oxygen at the time of film formation and forming a dense film, a stable film whose hygroscopicity does not increase even when fluorine is added can be obtained.

For example, when growing a SiOF film by a CVD method using a mixed gas of SiH ₄ / N ₂ O / SiF ₄ ,
The gas flow ratio governs the film density most. Further, at present, film forming conditions are delicate and it is difficult to perform sufficient control. Therefore, if the film density is accurately measured using the measurement technique of the present embodiment, and the measurement result is fed back to the film forming conditions, the film quality control of the SiOF film can be managed with high accuracy.

As described above, according to the present embodiment, when calculating the average reflectance curve R _ave , the power function of the incident angle representing the attenuation of the reflectance when there is no unevenness on the sample surface, and the influence of the surface unevenness. The data is fitted using an analytical expression consisting of a term consisting of the product of the exponential function and the constant term representing the background. The set degree of freedom can be restricted in a correct manner.

Thus, even when the structure of the film is different, the analytical formula can be used with versatility and the measurement can be performed with high accuracy. In addition, the speed of measurement can be improved. In the present embodiment, the Debye-Waller type roughness is assumed as the term of the exponential function representing the influence of the surface unevenness, but an accurate expression may represent the attenuation of the reflectance due to the roughness. For example, an expression represented by exp (−q · q ′ · σ ² ) can be applied. Where q
And q ′ are momentum transfer vectors on and in the substrate, respectively, and are expressed as q = (4π / λ) √ (n ² −cos ² θ), and n is a complex refractive index.

Further, in the present embodiment, the curve R is fitted to the analytical expression by the least squares method, but another fitting method may be used. For example, a simplex method, a Marquardt method, or a maximum entropy method can be applied. Further, in the present embodiment, an analytical expression including the background constant B ₀ is presented. However, when the background component is small, fitting may be performed using an analytical expression not including the background constant B ₀ .

[0074]

As described above, according to the present invention, the film thickness of the film is measured by measuring the X-ray reflectance from the film, extracting interference vibrations from the X-ray reflectance, and performing Fourier transform of the interference vibrations. In the film thickness measurement method to be calculated, the term consisting of the product of the power function of the incident angle representing the attenuation of the reflectance when the film surface has no irregularities and the exponential function representing the influence of the film surface irregularities is used as the background. The average reflectance is obtained by fitting the X-ray reflectance to an analytical expression to which a constant term representing
Since the interference vibration is extracted using the X-ray reflectivity and the average reflectivity, the analytical formula incorporates the property of the reflectivity, and the degree of freedom that was conventionally set arbitrarily must be limited in a correct form. Can be. Accordingly, even when the structures of the films are different, the analytical formula can be used with versatility and the measurement can be performed with high accuracy. In addition, the speed of measurement can be improved.

Further, according to the present invention, physical properties of the film such as film thickness, film density, surface irregularities, etc. can be measured simply and accurately, and these measured values can be used immediately after film formation. . Therefore, it is possible to control the film formation process of the film, which has been impossible in the past, with high accuracy. The following methods can be used for the management and control. For example, at the beginning of a series of processes for forming a film under the same conditions, a sample for condition inspection is prepared and inspected to confirm and readjust the film forming conditions. In addition, the relationship between the physical property values of these films and the film forming conditions is created in advance as a calibration curve, and the measurement results are compared with the calibration curve to quickly adapt to the film forming conditions. Similarly, it is also very effective in the inspection of the film forming process through the sampling inspection. Thus, various thin films such as an oxide film, a metal film for an electrode, a metal film for a wiring, and a dielectric film in a large-scale semiconductor device, and a film which has been impossible in the technical field such as a metal thin film in a magnetic head. It is possible to control the thickness, film density, and surface irregularities, and through these controls, it is possible to control the quality of the product. Therefore, it greatly contributes to improvement of production efficiency and yield.

[Brief description of the drawings]

FIG. 1 is a diagram schematically illustrating a film thickness measuring method according to a first embodiment of the present invention.

FIG. 2 is a graph showing a result obtained by analyzing a reflectance curve obtained by simulation by a film thickness measuring method according to a first embodiment of the present invention.

FIG. 3 is a graph showing a result of analyzing a reflectance curve obtained by a simulation by a conventional film thickness measuring method.

FIG. 4 is a graph showing a result of analyzing a multilayer film (TiN / Ti / SiO ₂ / Si) by the film thickness measuring method according to the first embodiment of the present invention.

FIG. 5 is a graph showing a result of an analysis performed in FIG. 4 with an angle measurement region limited to a low angle side.

FIG. 6 is a graph showing an analysis result obtained by limiting an angle measurement region to a high angle side in FIG. 4;

FIG. 7 is a graph showing a result of analyzing a silicon oxide film by the film thickness measuring method according to the first embodiment of the present invention.

FIG. 8 is a graph showing a result of analyzing a titanium nitride film by the film thickness measuring method according to the first embodiment of the present invention.

FIG. 9 is a graph showing a result of analyzing a multilayer film (TiN / Ti / SiO ₂ / Si) by the film thickness measuring method according to the first embodiment of the present invention.

FIG. 10 is a graph showing a result of an analysis performed in FIG. 9 with an incident angle region limited to a range of 0.6 to 2 °.

FIG. 11 is a graph showing the results of analysis performed in FIG. 9 with the incident angle region limited to the range of 2 to 4 °.

FIG. 12 is a graph showing a result of an analysis performed in FIG. 9 with an incident angle region limited to a range of 4 to 6 °.

FIG. 13 is a graph showing a result of an analysis performed in FIG. 9 with the incident angle range set to the whole range of 0 to 6 °.

[Explanation of symbols]

R: X-ray reflectance curve R _ave : average reflectance curve Δ (θ): vibration component curve

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification code FI H01L 21/302 H01L 21/66 P 21/66 H05K 3/00 V H05K 3/00 H01L 21/302 Z (56) References Special features JP-A-8-254509 (JP, A) JP-A-6-221841 (JP, A) JP-A-4-340407 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G01B 15 / 00-15/08 G01N 23/00-23/227 G01R 23/16 H01L 21/203 H01L 21/205 H01L 21/302 H01L 21/66 H05K 3/00 B01J 19/00

Claims (9)

(57) [Claims] 1. A film thickness measuring method for measuring an X-ray reflectance from a film, extracting interference vibration from the measured X-ray reflectance, and calculating a film thickness of the film by Fourier transforming the interference vibration. In the term consisting of the product of the exponential function representing the influence of the surface irregularities of the film and the exponential function representing the effect of the surface irregularities of the film, the exponential function representing the effect of the surface irregularities of the film, and a power function of the incident angle representing the attenuation of the reflectance when the surface of the film has no irregularities. Calculating an average reflectance by fitting the measured X-ray reflectivity to an analytical expression to which the interference vibration is added, and extracting the interference vibration using the measured X-ray reflectivity and the average reflectance. Film thickness measurement method. 2. The method of measuring a film thickness according to claim 1, wherein the analytic formulas are: intensity I ₀ , incident angle of X-ray θ, θ, origin correction value of the incident angle θ ₀ , roughness of the film surface. Is represented by I ₀ (θ−θ ₀ ) ⁻⁴ · exp の − (4πσ / λ · sin θ) ² ｝ + B ₀ , where σ is the mean square value of X, λ is the X-ray wavelength, and B ₀ is the background constant. A film thickness measuring method, characterized in that: 3. The method of measuring a film thickness according to claim 2 , wherein a square mean value of the roughness obtained by fitting the measured X-ray reflectance to the analytical expression is
A method for measuring a film thickness, characterized by estimating unevenness of the film surface. 4. The film thickness measuring method according to claim 1, wherein the interference vibration is obtained by dividing the measured X-ray reflectance by the average reflectance. Thickness measurement method. 5. The method for measuring a film thickness according to claim 1, wherein the film is a multilayer film having two or more layers, and a layer located on the surface side of the film. The angle of incidence is low, and the range of the angle of incidence is limited based on the position of the layer to be measured in the multilayer film so that the angle of incidence is high for a layer located on the inner side of the film. A film thickness measuring method, characterized in that the film thickness is measured. 6. A method of manufacturing a film for forming a film on an underlying substrate, wherein the film thickness of the film is measured by the film thickness measuring method according to claim 1, wherein the film thickness is measured. A method for producing a film, comprising: adjusting a film forming condition of the film using a film thickness of the film. 7. A method for manufacturing a film for forming a film on an underlying substrate, comprising: measuring the film thickness of the film by the film thickness measuring method according to claim 1; Calculate the X-ray reflectivity with the film thickness as an initial value, by fitting the calculated X-ray reflectivity and the measured X-ray reflectivity, determine the film density of the film, at least one physical quantity of surface irregularities, A method for producing a film, comprising adjusting a film-forming condition of the film using at least one physical quantity of a film thickness, a film density, and surface irregularities. 8. The method for producing a film according to claim 7, wherein at least one physical quantity of the film thickness, the film density, and the surface irregularities is used to perform an oxidation treatment, a nitridation treatment, an oxynitridation treatment, a sputtering treatment, C
In a manufacturing apparatus that performs any one of VD processing, electron beam evaporation processing, ion beam evaporation processing, molecular beam evaporation processing, and liquid phase growth processing, at least one parameter of a film forming temperature, a gas introduction amount, and a gas mixing ratio is set. A method for producing a film, comprising adjusting. 9. A method of etching a film formed by etching at least a part of a film formed on a base substrate, wherein the film is etched by the film thickness measuring method according to claim 1. Description: Measuring the thickness of the film, and adjusting the etching conditions of the film using the measured film thickness of the film.