rus
Issue #7-8/2018
D.Petrakov, N.Gerasimenko, N.Medetov, D.Smirnov, R.Suyundukov
The analysis of structural properties and parameters of objectsusing an integrated approach to processing x-ray optical measurement
The analysis of structural properties and parameters of objectsusing an integrated approach to processing x-ray optical measurement
Investigations of thin-film microelectronic structures of low-k dielectrics and TiN diffusivebarrier layers using proposed complex of X-ray methods, particularly, reflectometry and diffused scattering have been accomplished. The advantages of this approach are that it allows of resolving ambiguities, such as "density-roughness" appeared at solving of inverse X-ray problems, identifying features of the studied structures formation and developing an analytical complex combined by self-consistent processing of the data obtained by several measurement methods in practice. In future, it is planned to adapt it to X-ray spectral and X-ray diffraction analysis methods to study the elemental and phase composition of objects, including the analysis of ore minerals.
Теги: software thin-film nanostructures x-ray optical methods комплекс рентгенооптических методов программное обеспечение тонкопленочные наноструктуры
To solve the problems arising in research and analysis of the structure and dimensional parameters of objects, there exist various measurement methods, such as X-ray reflectometry and refractometry, diffraction and X-ray fluorescence analysis, transmission electron microscopy, optical ellipsometry, etc. However, the most interesting method for measurements of nanoelectronic thin-film structures is X-ray reflectometry. This non-destructive method for investigations of thin-film samples permits to measure not only thickness of layers in a single-layer and multilayer structure, but also to study roughness of every layer surfaces and calculate their values of densities in these structures and to determines hidden and unaccounted sublayers [1].
APPLICATION OF THE INTEGRATED APPROACH TO STUDY OF
THIN-FILMS STRUCTURES
X-ray reflectometry is a standard method for studying multilayer thin-film structures, however, a study of only the mirror component of X-ray reflection does not allow to separate the contributions to the reflectometric picture made by a normal component of density gradient of a layer material and by roughnesses and non-homogeneity on surface boundaries. Thus, X-ray refractometry and diffuse X-rays scattering were included into the integrated approach together with a method of relative reflectometry [2].
The study of X-ray refraction in thin-film structures permits to directly calculate X-ray refractive index and, therefore, obtain information about density distribution of the material in the investigated structure [3]. To clarify the parameters of the interface boundary roughnesses, the diffuse X-ray scattering method has been applied. The proposed approach allows to fix the problem of "density-roughness" ambiguity when solving the inverse problem of reflectometry, which consists in the fact that it is possible to restore only one-dimensional integral profile of polarizability across the depth of the structure, which can be preset both by a gradient of the physical density of the deposited layers and by presence of interlayer roughness. Use of additional techniques for self-consistent data processing solves this problem and allows of getting more complete understanding of thin-film structures and their composition. The above approach was implemented in special software [4] which enables to organize a parallel experimental processing of the data obtained by several independent X-ray methods (Fig.1). Thus, the proposed approach means that reflectometry data are corrected using refractometry and diffusion scattering of X-ray methods, hence, we not only resolve ambiguities, but also obtain more detailed information about a structure and composition of thin films [5–6].
In future, it is also planned to expand applicability of this integrated approach to develop an automated complex for processing of X-ray measurement results and analyzing a wide range of objects from various areas of engineering and technology, including research of ore and mineral samples by complementary methods of X-ray analysis (for example, using X-ray structural analysis and spectroscopy of x-ray absorption) within a single measuring platform without using a wide range of measuring equipment. Such system permits to get not only more detailed information about qualitative and quantitative elemental composition and phase of studied objects, but also significantly improve accuracy of the analysis and accelerate processing of obtained results, which will reduce the research cost.
Research of thin-film structures using integrated approach for X-ray optical data treatment. Theoretical structures of low-k dielectric of SiCN (92 nm) and SiOC (180 nm) and diffusive-barrier layers based on TiN (5 nm) have been investigated. The studied samples have been produced at JSC "Angstrem-T" plant (Zelenograd). These structures were chosen for the study because the X-ray reflectometry method only was not sufficient to characterize the structure parameters due to presence of unaccounted sublayers and existence of a density gradient of the material, which led to appearance of the "density-roughness" ambiguity when solving the inverse problem of this research method. Experimental measurements of the structures were made by CompleXRay analytical system. For the first time the mode of relative measurements was realized by measuring a signal ratio in two or more selected spectral lines [7]. This mode excludes hardware faults of X-ray measurements and allows to make express measurements at near-zero scattering angles. Besides, thin-film low-k SiCN/SiOC dielectrics were studied [8–9] using UV-ellipsometry, and TiN structure of diffusive-barrier layer was investigated by a destructive method of transmission electronic microscopy, accordingly. However, as it was described above, X-ray methods ensure not only a higher measurement accuracy but are, also, non-destructive and, when used in combination, enable to obtain reliable results in a minimum time.
THE RESULTS OF LOW-K SICN/SIOC DIELECTRICS STRUCTURE INVESTIGATIONS
In order to correctly determine the basic parameters of the studied structure, it was not sufficient to use the X-ray reflectometry method only because of probable presence of SiCN and SiOC layers density gradient. In order to solve this problem the refractometry method has been used. The refractogram of a structure shown in Fig.2 indicates that the angular positions of the refraction maxima are 0.142° and 0.176°. As a results, the values of refraction index are δ1 = 3.98 and δ2 = 7.92 for every layer and correspond to the density values of ρSiCN = 1.2 and ρSiOC = 2,3 g/sm3. Fig.3 presents a reflectometry dependence obtained by a complex use of reflectometry and refractometry methods. The measurement results are summarized in Table 1.
As it is shown in Table 1 and Fig.3 the low-k dielectric layer density is equal to 1.3±0.2 g/sm3 that means a presence of density gradient in a sample. Thus, the method of refractometry determined the density of the layers with high accuracy for the given structure which proves efficiency of such complex approach implementation.
THE RESULTS OF TIN DIFFUSIVE-BARRIER INVESTIGATIONS
While depositing TiN, some problems frequently arise because of complexity of stoichiometric titanium nitride layers formation. Nitride films are deposited from a gas phase of metalorganic precursor in the course of a multistage technological process followed by condensing in a glow-discharge plasma thereby creating several sublayers of different density and stoichiometry. The X-ray reflectometry method allows to calculate sublayer densities but it is difficult to use this method for analyzing the structure because of the arising "density-roughness" ambiguity. In this paper we propose to use the X-ray diffusive scattering method to solve the ambiguity problem. Fig. 4 and 5 present reflectogram of diffusive-barrier TiN layer and diffusive scattering curve, correspondingly. The results of reflectometry and diffusive scattering are shown in Table 2.
When summarizing the data on the phase composition, density, roughness and sublayers in the TiN film were identified. A thin amorphous TiNOx sublayer of 3.1 g/sm3 density on the surface of the diffusive-barrier surface is, actually, TiNOx oxygenized in air because of the absence of the top protective layer and presence of a residual charge near the surface area after treatment in plasma. Thin amorphous TiNy sublayer of 2.2 g/sm3 density that has been formed at a boundary with a silicon dioxide layer presents a loose film which was not completely condensed and crystallized during the treatment in a plasma media. The X-ray diffusive scattering data proves that the sample roughness is very low and that appearance of the reflectometry curve depends on presence of several unaccounted sublayers in the investigated structure.
CONCLUSIONS
As the results of the study indicated, the analysis of basic parameters of thin films using a single method may not provide reliable information about a structure of the object and the subsequent adjustment of these parameters in accordance with the requirements of the process.
It happens due to the following factors:
accuracy of the applied method and physical limitations when studying the object do not allow of using a particular method in all cases;
usage of the chosen method may lead to different problems connected with peculiarities of it application, for example, in case of X-ray reflectometry they are "density-roughness" ambiguities, for refractometry – it is limitation of layers measurement when the layer thickness does not exceed few tens of nanometers. The majority of problems described above appear during investigations of nano-scale structures or when such kinds of objects are located on thin-film samples.
In future it is planned to adapt the described self-consistent approach to X-ray spectral and X-ray structure analyses when studying the elemental and phase composition of objects, including the analysis of ore minerals. It will enable to get more complete and detailed picture of the studied objects and, besides, raise a sensitivity threshold when detecting individual elements that cannot be identified when using individual analytical methods. ■
The research was funded out of the grant for performance of scientific and (or) scientific and technological projects in 2018–2020 in Kazakhstan (project No. AP05133354).
APPLICATION OF THE INTEGRATED APPROACH TO STUDY OF
THIN-FILMS STRUCTURES
X-ray reflectometry is a standard method for studying multilayer thin-film structures, however, a study of only the mirror component of X-ray reflection does not allow to separate the contributions to the reflectometric picture made by a normal component of density gradient of a layer material and by roughnesses and non-homogeneity on surface boundaries. Thus, X-ray refractometry and diffuse X-rays scattering were included into the integrated approach together with a method of relative reflectometry [2].
The study of X-ray refraction in thin-film structures permits to directly calculate X-ray refractive index and, therefore, obtain information about density distribution of the material in the investigated structure [3]. To clarify the parameters of the interface boundary roughnesses, the diffuse X-ray scattering method has been applied. The proposed approach allows to fix the problem of "density-roughness" ambiguity when solving the inverse problem of reflectometry, which consists in the fact that it is possible to restore only one-dimensional integral profile of polarizability across the depth of the structure, which can be preset both by a gradient of the physical density of the deposited layers and by presence of interlayer roughness. Use of additional techniques for self-consistent data processing solves this problem and allows of getting more complete understanding of thin-film structures and their composition. The above approach was implemented in special software [4] which enables to organize a parallel experimental processing of the data obtained by several independent X-ray methods (Fig.1). Thus, the proposed approach means that reflectometry data are corrected using refractometry and diffusion scattering of X-ray methods, hence, we not only resolve ambiguities, but also obtain more detailed information about a structure and composition of thin films [5–6].
In future, it is also planned to expand applicability of this integrated approach to develop an automated complex for processing of X-ray measurement results and analyzing a wide range of objects from various areas of engineering and technology, including research of ore and mineral samples by complementary methods of X-ray analysis (for example, using X-ray structural analysis and spectroscopy of x-ray absorption) within a single measuring platform without using a wide range of measuring equipment. Such system permits to get not only more detailed information about qualitative and quantitative elemental composition and phase of studied objects, but also significantly improve accuracy of the analysis and accelerate processing of obtained results, which will reduce the research cost.
Research of thin-film structures using integrated approach for X-ray optical data treatment. Theoretical structures of low-k dielectric of SiCN (92 nm) and SiOC (180 nm) and diffusive-barrier layers based on TiN (5 nm) have been investigated. The studied samples have been produced at JSC "Angstrem-T" plant (Zelenograd). These structures were chosen for the study because the X-ray reflectometry method only was not sufficient to characterize the structure parameters due to presence of unaccounted sublayers and existence of a density gradient of the material, which led to appearance of the "density-roughness" ambiguity when solving the inverse problem of this research method. Experimental measurements of the structures were made by CompleXRay analytical system. For the first time the mode of relative measurements was realized by measuring a signal ratio in two or more selected spectral lines [7]. This mode excludes hardware faults of X-ray measurements and allows to make express measurements at near-zero scattering angles. Besides, thin-film low-k SiCN/SiOC dielectrics were studied [8–9] using UV-ellipsometry, and TiN structure of diffusive-barrier layer was investigated by a destructive method of transmission electronic microscopy, accordingly. However, as it was described above, X-ray methods ensure not only a higher measurement accuracy but are, also, non-destructive and, when used in combination, enable to obtain reliable results in a minimum time.
THE RESULTS OF LOW-K SICN/SIOC DIELECTRICS STRUCTURE INVESTIGATIONS
In order to correctly determine the basic parameters of the studied structure, it was not sufficient to use the X-ray reflectometry method only because of probable presence of SiCN and SiOC layers density gradient. In order to solve this problem the refractometry method has been used. The refractogram of a structure shown in Fig.2 indicates that the angular positions of the refraction maxima are 0.142° and 0.176°. As a results, the values of refraction index are δ1 = 3.98 and δ2 = 7.92 for every layer and correspond to the density values of ρSiCN = 1.2 and ρSiOC = 2,3 g/sm3. Fig.3 presents a reflectometry dependence obtained by a complex use of reflectometry and refractometry methods. The measurement results are summarized in Table 1.
As it is shown in Table 1 and Fig.3 the low-k dielectric layer density is equal to 1.3±0.2 g/sm3 that means a presence of density gradient in a sample. Thus, the method of refractometry determined the density of the layers with high accuracy for the given structure which proves efficiency of such complex approach implementation.
THE RESULTS OF TIN DIFFUSIVE-BARRIER INVESTIGATIONS
While depositing TiN, some problems frequently arise because of complexity of stoichiometric titanium nitride layers formation. Nitride films are deposited from a gas phase of metalorganic precursor in the course of a multistage technological process followed by condensing in a glow-discharge plasma thereby creating several sublayers of different density and stoichiometry. The X-ray reflectometry method allows to calculate sublayer densities but it is difficult to use this method for analyzing the structure because of the arising "density-roughness" ambiguity. In this paper we propose to use the X-ray diffusive scattering method to solve the ambiguity problem. Fig. 4 and 5 present reflectogram of diffusive-barrier TiN layer and diffusive scattering curve, correspondingly. The results of reflectometry and diffusive scattering are shown in Table 2.
When summarizing the data on the phase composition, density, roughness and sublayers in the TiN film were identified. A thin amorphous TiNOx sublayer of 3.1 g/sm3 density on the surface of the diffusive-barrier surface is, actually, TiNOx oxygenized in air because of the absence of the top protective layer and presence of a residual charge near the surface area after treatment in plasma. Thin amorphous TiNy sublayer of 2.2 g/sm3 density that has been formed at a boundary with a silicon dioxide layer presents a loose film which was not completely condensed and crystallized during the treatment in a plasma media. The X-ray diffusive scattering data proves that the sample roughness is very low and that appearance of the reflectometry curve depends on presence of several unaccounted sublayers in the investigated structure.
CONCLUSIONS
As the results of the study indicated, the analysis of basic parameters of thin films using a single method may not provide reliable information about a structure of the object and the subsequent adjustment of these parameters in accordance with the requirements of the process.
It happens due to the following factors:
accuracy of the applied method and physical limitations when studying the object do not allow of using a particular method in all cases;
usage of the chosen method may lead to different problems connected with peculiarities of it application, for example, in case of X-ray reflectometry they are "density-roughness" ambiguities, for refractometry – it is limitation of layers measurement when the layer thickness does not exceed few tens of nanometers. The majority of problems described above appear during investigations of nano-scale structures or when such kinds of objects are located on thin-film samples.
In future it is planned to adapt the described self-consistent approach to X-ray spectral and X-ray structure analyses when studying the elemental and phase composition of objects, including the analysis of ore minerals. It will enable to get more complete and detailed picture of the studied objects and, besides, raise a sensitivity threshold when detecting individual elements that cannot be identified when using individual analytical methods. ■
The research was funded out of the grant for performance of scientific and (or) scientific and technological projects in 2018–2020 in Kazakhstan (project No. AP05133354).
Readers feedback
