JP2000035408A - Film structure analysis method using x-ray reflectance method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To qualitatively grasp structure analysis or interface condition in a short time by Fourier transforming the vibration component of reflected X-ray intensity obtained by an X-ray reflectance method, and analysing film construction from a Fourier transformation spectrum. SOLUTION: X-rays are radiated on a multiple layer thin film specimen to scan in θ-2θ of small angle θ (usually θ<=10 degrees), the vibration component of obtained reflected X-ray intensity is Fourier transformed, and Fourier transformation spectrum as the function of film thickness is obtained. In this Fourier transformation spectrum, total film thickness of the specimen is obtained from the peak existing on the film thickness position of high peak intensity and the thickest. Further, based on this spectrum, the film thickness of respective constituting thin film layers and the accumulated film thickness can be obtained. Further, by X-ray reflectance analysis in which the film thickness of the respective obtained constituting layers are set as initial values, the film thickness, density and interface roughness (interface condition) of the respective constituting layers can be qualitatively obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for analyzing a film structure using an X-ray reflectivity method.
X-rays are incident at a small incident angle θ of about 0 °, and the film structure analysis using the X-ray reflectivity method for analyzing the structure of the multilayer thin film sample from the change in the X-ray reflectivity with respect to the change in the incident angle of the X-rays It is about the method.

[0002]

2. Description of the Related Art Conventionally, an X-ray diffractometer or an X-ray reflectivity method has been used as a method for analyzing a film structure or a crystal structure of a sample using X-rays. It is used as an effective method for evaluating the surface structure of a sample having a multilayer structure in which various thin films such as a semiconductor device and an MR (magnetoresistive effect) element are deposited.

Recently, since a multilayer thin film in which a plurality of different layers are stacked on a magnetic disk and an MR head constituting a main portion of a hard disk of a computer has been used, a multilayer thin film has been developed in order to improve a production yield. The thickness of each layer constituting the thin film and the interface roughness between each layer (rou
It is required to perform film structure analysis including ghness) with higher accuracy.

The X-ray reflectivity method is a method of evaluating the physical properties by matching the dependence of the reflected X-ray intensity profile on the X-ray incident angle on a multilayer thin film sample with the simulation results. In the case of a flat sample, the reflected X-ray intensity theoretically attenuates in inverse proportion to the fourth power of the incident angle θ of the X-ray to the sample, and more rapidly when the thin film / thin film interface is not flat. .

Here, a schematic configuration of an X-ray reflectivity measuring apparatus used for the X-ray reflectivity method will be described with reference to FIG. See FIG. 6. The X-ray reflectometer measures the X-ray source 11
A slit 12 for shaping an incident X-ray 20 from the source 11,
A channel-cut crystal 13 made of a pair of parallel plate-like silicon crystals that suppress the angular divergence of the incident X-ray 20 that has passed through the slit 12. Slit 1
4. An incident X-ray intensity monitor 15 comprising an ion chamber for measuring the intensity of the incident X-rays 20 passing through the slit 14, a sample 16 to be measured placed thereon, and an X-ray incident angle θ with respect to the sample 16 to be measured is arbitrarily set. X-ray detector 19 comprising a goniometer 17, a slit 18 for limiting the incidence of reflected X-rays 21 from the sample 16 to be measured, and a scintillation counter for detecting the reflected X-rays 21 through the slit 18.
It consists of.

The incident X-rays 20 are reflected at a reflectance corresponding to the film structure of the sample 16 to be measured, and the reflected X-rays 21 are reflected.
Is detected by the X-ray detector 19 through the slit 18. In this case, the angle of incidence θ of the sample 16 to be measured with respect to the incident X-ray 20 by the goniometer 17, and thus the rotation angle of the goniometer is changed. The measurement will be performed while doing so.

In this case, the slit 18 and the X-ray detector 1
Numeral 9 is attached to an arm (not shown) fixed to the goniometer 17 so as to rotate in conjunction with the rotation of the goniometer 17.
The operation is set so that the line incident angle θ and the elevation angle θ ′ of the X-ray detector 19 become equal, and the axial direction of the X-ray detector 19 is always set to a relationship of 2θ with respect to the incident X-ray 20. Therefore, such a method is called θ-2θ scan.

In the measurement data measured by the X-ray reflectivity method, as described above, the intensity of the reflected X-ray 21 is theoretically inversely related to the fourth power of the X-ray incident angle θ to the sample. Since there is an incident angle θ dependence that attenuates proportionally, in order to cancel the effect of the incident angle θ, the baseline is determined using the least squares method, and only the vibration component included in the measurement data is extracted. I do.

Next, a simulation result obtained by appropriately changing values of the film thickness, density, and interface roughness of each film, which are parameters in the analysis model, and a vibration component included in the measurement data are compared with each other to obtain a predetermined value. The thickness of each layer and the like are determined by performing a least-squares fitting method so as to be within an error.

[0010]

However, when analyzing the film structure of the MR element using such an X-ray reflectivity method, when the film thickness or the like is obtained by the least squares fitting method, the analysis result is not a predetermined value. However, there is a problem that it takes a long time before the error falls within the range.

[0011] Further, the variation of the MR characteristics due to the film formation conditions during the manufacturing of the MR element or the deterioration of the MR characteristics caused by the heat treatment process is a problem related to the reliability of the product. It is considered that the deterioration of the MR characteristics is caused by a state change due to diffusion or the like at the lamination interface.

However, when analyzing a film structure in which the interface state has changed, it cannot be distinguished from an interface roughness between two layers and a diffusion layer formed by interdiffusion between the two layers. Due to the problem, the analysis result depends on the analysis model, and there is a problem that there are multiple analysis solutions even in the same analysis model, and it is difficult to determine which is the true solution .

Accordingly, an object of the present invention is to make it possible to perform a structural analysis of a sample having an unknown structure in a short time and to qualitatively grasp a change in an interface state when the X-ray reflectivity method is used. And

[0014]

FIG. 1 is an explanatory view of the principle configuration showing the procedure of the present invention. Referring to FIG. 1, means for solving the problem in the present invention will be described. See FIG. 1 (1) In the present invention, in a film structure analysis method using an X-ray reflectivity method, a small angle θ-2θ scan is performed on a multilayer thin film sample, and a vibration component of the obtained reflected X-ray intensity is obtained. Is subjected to a Fourier transform, and the film structure of the multilayer thin film sample is analyzed from the Fourier transform spectrum.

As described above, first, the reflectance is measured by the X-ray reflectance method in the same manner as in the prior art, and the vibration component is extracted from the measured data. Analysis of the film structure of a multilayer thin film sample from the peak position and intensity in the Fourier transform spectrum in a shorter time than the analysis using the least squares fitting method, for example, determination of the film thickness from the Fourier transform spectrum or qualification of interface roughness Analysis can be performed. Note that the small angle θ in this case is usually θ ≦
It is about 10 °.

(2) Further, according to the present invention, in the above (1), the total film thickness of the multilayer thin film sample is obtained from a peak having a strong peak intensity in the Fourier transform spectrum and present at the position of the thickest film thickness. It is characterized by.

As described above, by obtaining the Fourier transform spectrum as a function of the film thickness by performing the Fourier transform on the measurement result of the small angle θ-2θ scan, the peak intensity in the Fourier transform spectrum is strongest and the thickest. The total film thickness of the multilayer thin film sample can be obtained with considerable accuracy from the peak existing at the film thickness position.

(3) Further, according to the present invention, in the above (1), the film thickness of each constituent layer constituting the multilayer thin film sample and the film thickness obtained by laminating each constituent layer are determined based on the Fourier transform spectrum. Features.

As described above, since the peak in the Fourier transform spectrum is obtained corresponding to the interface between the constituent layers,
The film thickness of each constituent layer or the film thickness of each constituent layer, that is, the cumulative film thickness, can be obtained from the position of this peak.

(4) According to the present invention, in the above (3), the film thickness of each constituent layer constituting the multilayer thin film sample and the film thickness obtained by laminating each constituent layer are determined by the interval between peaks in the Fourier transform spectrum. It is characterized by.

As described above, since the peak in the Fourier transform spectrum is obtained corresponding to the interface between the constituent layers,
From the interval between the peaks, the individual film thickness of each constituent layer, or the film thickness of the stacked constituent layers, that is, the total film thickness of a plurality of constituent layers can be obtained.

(5) In the present invention, in the above (4), X-ray reflectivity analysis is performed using the obtained film thickness of each constituent layer as an initial value, and the film thickness, density, and And the interface roughness is determined.

As described above, X-ray reflectivity analysis is performed using the film thickness of each constituent layer obtained from the Fourier transform spectrum as an initial value, that is, the film thickness, density, and interface roughness of each constituent layer are used as parameters. By performing the analysis by the least squares fitting, it is possible to converge within a predetermined error in a shorter time, and the film thickness, density, and density of each constituent layer constituting the unknown multilayer thin film sample by the X-ray reflectivity analysis In addition, the film structure such as interface roughness can be analyzed in a shorter time.

(6) Further, according to the present invention, in the above (1), the position of each peak in the Fourier transform spectrum,
It is characterized in that the internal interface state of the multilayer thin film sample is evaluated based on the change in the strength and the half width.

As described above, the position, intensity, and half width of each peak in the Fourier transform spectrum are determined by the internal interface state of the multilayer thin film sample, particularly, the interface state converted by the film forming condition or the heat treatment condition, for example, the interface roughness. Therefore, the internal interface state of the multilayer thin film sample can be easily and qualitatively evaluated from the position, intensity, and half width of each peak in the Fourier transform spectrum.

[0026]

FIG. 6 and FIGS. 2 to 4 show the procedure of a film structure analysis method using the X-ray reflectivity method according to the first embodiment of the present invention. explain. First, although not shown, a laminated structure of a spin-valve magnetoresistive element used as a multilayer thin film sample to be measured will be described. A thickness is formed by sputtering on a silicon substrate having a (100) plane as a main surface. Underlayer consisting of 5 nm Ta film and 4 nm thick NiFe film, 2.2 nm thick Co
_A free layer made of ₉₀ Fe ₁₀ film, an intermediate layer made of Cu film,
A pinned layer made of a Co ₉₀ Fe ₁₀ film having a thickness of 2.2 nm, an antiferromagnetic material layer made of a Pd ₃₂ Pt ₁₇ Mn ₅₁ (PPM) film having a thickness of 24, and a Ta film having a thickness of 6 nm are sequentially laminated. ,
Ta / NiFe / CoFe / Cu / CoFe / PPM /
A multilayer thin film structure made of Ta is formed. When an actual spin-valve magnetoresistive element is formed, 30O
The deposition is performed while applying a magnetic field of about e (磁 界 2.4 kA / m).

In this case, the thickness x of the Cu film constituting the intermediate layer is x = 2.1 nm, 2.5 nm, 3.0 nm, 4.
Four spin-valve magnetoresistive elements having four settings of 0 nm are prepared. Here, the film thickness is a design value, and the deposition time is controlled based on the deposition rate at the time of forming each layer so that the film thickness of the design value is obtained.

Referring again to FIG. 6, the X-ray reflectivity of the obtained four spin-valve magnetoresistive elements is measured by the X-ray reflectivity method. This is the same as the X-ray reflectivity measuring apparatus shown in FIG. 6, and uses four spin-valve magnetoresistive elements having the above-described Cu films having different thicknesses as the sample 16 to be measured. The measurement is performed by irradiating incident X-rays 20 having a single wavelength of .62 °.

Referring to FIG. 2, the dotted line in FIG. 2 shows an example of the measurement result of the X-ray reflectivity by θ-2θ scan, and the period of the obtained vibration component reflects the film thickness of the sample to be measured. The vibration with a small cycle corresponds to the total thickness of the sample to be measured, and the vibration with a large cycle corresponds to the thickness of each thin film constituting the sample to be measured. The relationship is inversely proportional.

The measurement data on the X-ray reflection intensity indicated by the dotted line indicates that, as described above, the intensity of the reflected X-ray 21 is theoretically attenuated in inverse proportion to the fourth power of the X-ray incident angle θ to the sample. Since there is an incident angle θ dependency, in order to cancel out the effect of the incident angle θ dependence, it is necessary to determine a baseline using the least squares method and extract only the vibration component included in the measurement data, The result of extracting only the vibration component is shown by the solid line in FIG.

Next, the measurement data obtained by extracting only the vibration component is subjected to Fourier transform using the film thickness d as a function to obtain a Fourier transform spectrum. Actually, measurement data is input to a computer using Fourier transform software corresponding to X-ray reflectance analysis, and calculation is automatically performed.

FIG. 3 shows the Fourier transform spectrum of the X-ray reflection intensity data obtained in this manner. The Fourier transform spectrum corresponding to the four film thicknesses of the Cu film is obtained by dividing the peak intensity into arbitrary units. Are displayed while shifting. Since there is a predetermined relationship between the film thickness d and the X-ray incident angle θ, the peak position of the Fourier transform spectrum is displayed on the horizontal axis in units of the film thickness (nm).

The position of each peak of each Fourier transform spectrum reflects the position of the interface between the layers constituting the spin valve magnetoresistive element, and such a position of the peak causes the interface from the surface of the silicon substrate to the surface. The position can be determined. In the drawing, Ta ₂ O ₅ on the left side, that is, on the silicon substrate side, and the third Ta / T
The peak of a ₂ O _{5 is} a peak corresponding to a natural oxide film of Ta formed by O ₂ contained in the atmosphere in the process of depositing a Ta film as an underlayer.

Among these peaks, the highest peak on the right side, that is, the peak with the highest intensity on the right side is the total film thickness (To
tal thickness),
Also in this case, a natural oxide film Ta ₂ O ₅ is formed on the surface of the Ta film.
Are formed. Although a peak is also seen on the right side of the high peak representing the total film thickness, the reflection intensity on the surface of the sample generally becomes very high, so the highest peak on the right side is the peak corresponding to the total film thickness. The peak on the right side is empirically determined to be a peak due to a higher-order component.

As is apparent from FIG. 3, as the thickness of the Cu layer increases, the Ta /
The peak corresponding to NiFe / CoFe / Cu shifts to the right, and it becomes possible to analyze the film thickness distribution of the multilayer thin film structure using such a Fourier transform spectrum.

FIG. 4 shows the result of measuring the film thickness of each film from such a Fourier transform spectrum. As shown in FIG. 0.7-4.8n
m, and the NiFe film having a design value of 4 nm
The CoFe film as a free layer having a thickness of 4.4 nm and a design value of 2.2 nm is 1.5 to 1.6 nm, and a Cu film having a design value of 2.1 nm, 2.5 nm, 3.0 nm and 4.0 nm. Are 2.6 nm, 2.9 nm, and 3.2 n, respectively.
m, 4.3 nm, a CoFe film as a pinned layer having a design value of 2.2 nm is 1.7 to 1.9 nm, a PPM film having a design value of 24 nm is 23 nm, and a Ta film having a design value of 6 nm. Is 4.7 to 4.8 nm, and Ta
The native oxide film Ta ₂ O ₅ formed on the film is 1.7 to
2.1 nm.

In this case, the actual film thickness of each film in each sample does not always agree with the design value due to subtle differences in the conditions at the time of film formation of each sample. It can be seen that the thickness of the film can be analyzed with considerable accuracy only by performing the Fourier transform on the vibration component.

Next, in order to perform a more accurate analysis using such a Fourier transform spectrum, the film thickness data obtained from the Fourier transform spectrum is used as an initial value of the film thickness, and an analysis similar to the conventional one is performed. Using the model, the least squares fitting was performed using the film thickness, density, and interface roughness as parameters.
The analysis was performed such that the fitting was terminated when the value became 0% or less, and the film thickness at that time is also shown in FIG.

Referring again to FIG. 4, as is apparent from FIG. 4, when the least squares fitting is performed, a value closer to the design value than the value obtained from the Fourier transform spectrum is obtained. By combining with the least squares fitting, it is possible to analyze the film structure with higher accuracy.

In this case, as described above, when performing the least squares fitting, the film thickness data obtained from the Fourier transform spectrum is used as the initial value of the film thickness.
The time required to converge to the predetermined fitting accuracy R is shorter than before, and the throughput is improved. The fitting accuracy R when the thickness of the Cu film is 2.1 nm is 1.64%.
When the thickness of the Cu film is 2.5 nm, the fitting accuracy R is 1.54%, and the thickness of the Cu film is 3.0 n.
m, the fitting accuracy R is 1.83%, and C
The fitting precision R when the thickness of the u film is 4.0 nm is 1.79%, and the smaller the fitting precision R, the higher the precision.

Next, a second embodiment of the present invention relating to a qualitative analysis method of interface roughness will be described with reference to FIG. FIG. 5 shows a spin-valve magnetoresistive element having a multi-layered thin film structure as in the case of the first embodiment, and measurement was performed using the X-ray reflectometer shown in FIG. 9 shows a Fourier transform spectrum obtained by extracting only a vibration component from the X-ray reflection intensity measurement data and performing a Fourier transform.

In this case, however, the thickness of the Cu film was fixed at 2.5 nm instead of the thickness of the Cu film, and the interface roughness between the Cu film and the CoFe film serving as the pinned layer was 0.1 nm.
0.2 nm, 0.3 nm, 0.45 nm, 0.6 nm,
Samples in seven states of 0.7 nm and 0.8 nm were used. However, the design value of the thickness of each layer is slightly changed from that of the first embodiment, and the thickness of the PPM film is 25 nm.
The interface roughness of the interface of each film was 0.3 n as shown in the figure.
m, 0.4 nm, 0.25 nm, 0.35 nm, yn
m, 0.50 nm, 0.55 nm, 0.50 nm, and 0.60 nm. In this case, the thickness of the Ta film is 5 nm, but it is 6 nm as a design value, and is 5 nm because a part of the Ta film is oxidized by the surface oxidation and converted into a Ta ₂ O ₅ layer.

The interface roughness represents half the average value of the height of the convex portion and the average value of the height (depth) of the concave portion in the unevenness at the interface between the two films. In order to change the interface roughness, in the second embodiment, the temperature at the time of film formation is changed to change the size of the interface roughness.

As is clear from the figure, when the interface roughness is changed, the peak corresponding to Ta / NiFe / CoFe / Cu, which is connected by a solid line corresponding to the interface between the Cu film and the CoFe pinned layer, indicates that the interface roughness becomes larger. Therefore, the peak shifts to the right and the intensity of the peak decreases.

This is because when the thin film / thin film interface is not flat, that is, when the interface roughness is large, the angle of incidence locally changes due to the unevenness of the position where the X-ray enters.
This is because the line reflection intensity is rapidly attenuated, and it can be qualitatively understood that the lower the peak intensity, the greater the interface roughness.

More specifically, the half width of the peak also slightly increases as the interface roughness increases. This is because the interface position varies due to unevenness.
This is because the half width of the peak is large, and it can be qualitatively understood that the larger the half width, the larger the interface roughness.

Therefore, by using the Fourier transform spectrum, it is possible to qualitatively analyze the interface roughness, and hence the interface state, in a short time as compared with the conventional analysis of the interface roughness by the least squares fitting method.

Although the embodiments of the present invention have been described above, the present invention is not limited to the configurations described in the above embodiments, and various modifications are possible. For example, in the description of each of the above embodiments, a spin valve magnetoresistive element is used as a sample to be measured.
The film structure analysis method of the present invention is not limited to the analysis of the spin valve magnetoresistance effect element, but can be used for the film structure analysis of various multilayer thin film structures including a semiconductor device. Is particularly effective for analyzing thin films having a thickness of several to several tens nm (several hundreds of nm).

In the description of each of the above embodiments, only the analysis of the film thickness and the qualitative analysis of the interface roughness are described, but the film thickness obtained from such a Fourier transform spectrum is calculated by the least square method. X by fitting
By using the initial values for the line reflectance analysis, the absolute values of the density and the interface roughness, which are other parameters, can be easily obtained from the conventional X-ray reflectance analysis using the least squares fitting method.

In recent years, it has been attempted to use an Ag film instead of a Cu film as an intermediate layer for separating a free layer and a pinned layer constituting a spin valve magnetoresistive element. , Cu (density = 8.93 g / c
m ³ ) and Ag (density = 1.492 g / cm ³ ) have different densities, and accordingly, the structure factor for the X-ray wavelength is also different. Since different shapes and peaks with different intensities and half widths appear, for example, by stacking such data, it is also effective for qualitative analysis of unknown components of each thin film constituting the multilayer thin film structure. Can be demonstrated.

In the above-described second embodiment, measurement is performed using a sample in which the interface roughness is intentionally set by changing the film forming conditions. When it is qualitatively determined that the interface roughness is large by the Fourier transform spectrum in the sample under the conditions, the sample is prepared while changing the film formation condition or the heat treatment condition, and the film formation condition or the heat treatment condition to reduce the interface roughness is obtained. It becomes possible to do.

[0052]

According to the present invention, the X-ray reflectivity X
By simply performing a Fourier transform on the vibration component of the line reflection intensity measurement data, it is possible to analyze the film structure of the multilayer thin film structure in a short time without the need for the least squares fitting.
In addition, it is possible to qualitatively analyze changes in the interface structure that could not be modeled by the conventional analysis model. Furthermore, the analysis values obtained by the Fourier transform can be converted to the initial values of the X-ray reflectance analysis by the least squares fitting method. As a result, the film thickness and the like can be obtained more accurately and in a shorter time, and the accuracy and reliability of the material evaluation by the X-ray reflectivity method can be remarkably improved.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram of a basic configuration of the present invention.

FIG. 2 is an explanatory diagram of a reflectance measurement result according to the first embodiment of the present invention.

FIG. 3 is a Fourier transform spectrum of reflectance measurement data according to the first embodiment of the present invention.

FIG. 4 is an explanatory diagram of an analysis result according to the first embodiment of the present invention.

FIG. 5 is a Fourier transform spectrum relating to interface roughness according to the second embodiment of the present invention.

FIG. 6 is a schematic configuration diagram of an X-ray reflectance measuring device.

[Explanation of symbols]

 Reference Signs List 11 X-ray source 12 Slit 13 Channel cut crystal 14 Slit 15 Incident X-ray intensity monitor 16 Sample to be measured 17 Goniometer 18 Slit 19 X-ray detector 20 Incident X-ray 21 Reflected X-ray

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 2G001 AA01 BA18 CA01 DA02 EA01 FA09 FA12 FA14 FA18 FA29 FA30 GA01 GA13 HA01 JA08 KA01 KA11 KA12 LA11 LA20 MA05 NA10 NA11 NA12 NA13 NA15 NA17 PA12 RA10 SA02

Claims (6)

[Claims] In a method of analyzing a film structure using an X-ray reflectivity method, a multilayer thin film sample is irradiated with X-rays, a small angle θ-2θ scan is performed, and oscillation of the obtained reflected X-ray intensity is performed. X is characterized by performing a Fourier transform of the components and analyzing a film structure of the multilayer thin film sample from a Fourier transform spectrum.
A film structure analysis method using the line reflectance method. 2. In the Fourier transform spectrum,
2. The film structure analysis method using the X-ray reflectivity method according to claim 1, wherein a total film thickness of the multilayer thin film sample is obtained from a peak having a strong peak intensity and being present at a position of the thickest film thickness. 3. The X-ray reflection according to claim 1, wherein a thickness of each of the constituent layers constituting the multilayer thin film sample and a thickness of the stacked constituent layers are determined based on the Fourier transform spectrum. A film structure analysis method using the rate method. 4. The method according to claim 3, wherein the thickness of each of the constituent layers constituting the multilayer thin film sample and the thickness of each of the constituent layers are determined from the interval between the peaks in the Fourier transform spectrum. A film structure analysis method using the line reflectance method. 5. An X-ray reflectivity analysis is performed using the thickness of each of the constituent layers obtained as an initial value, and the film thickness of each of the constituent layers is determined.
The method for analyzing a film structure using the X-ray reflectivity method according to claim 4, wherein the density and the interface roughness are obtained. 6. Based on changes in the position, intensity, and half width of each peak in the Fourier transform spectrum,
2. The method according to claim 1, wherein an internal interface state of the multilayer thin film sample is evaluated.