Issue #6/2024
S.V.Sidorova, A.D.Kouptsov, O.V.Novikova, I.V.Kushnarev, A.A.Epikhin, E.E.Gusev
ASSESSMENT OF COATING THICKNESS EFFECT ON THE AMOUNT OF RESIDUAL MECHANICAL STRESSES IN AL2O3/Si
ASSESSMENT OF COATING THICKNESS EFFECT ON THE AMOUNT OF RESIDUAL MECHANICAL STRESSES IN AL2O3/Si
INTRODUCTION
Modern technological solutions in various areas of mechanical engineering, robotics, medicine and other industrial fields cannot be imagined without rapidly developing field of nanoengineering, which includes the thin-film coating formation technology.
Different types of thin films are distinguished according to their functional purpose and areas of use. For example, conductive layers are formed in printed circuit boards and power electronics for load transfer and switching between electrical parts [1, 2]. Insulating layers serve to form a barrier area beyond which no electrical power propagation takes place [3]. Thin films that enhance protective, antifriction, lubrication and corrosion-resistant properties of parts are used for deposition on the working surfaces of cutting tools, roll surfaces and equipment that is operated in aggressive oxygen-containing environments [4]. Optical thin film coatings – anti-reflective, fluorescent, reflective, etc., have found their application in solving everyday and design problems [5]. Also, decorative structures are actively used to create unique patterns and inscriptions [6].
It should be noted that practical implementation of nanoscale thin film coatings, which occurs due to the dimensional effect – the effect of enhancing the level of physical parameters of the material (electrical conductivity, magnetic, optical and mechanical properties) due to predominance of the surface area over the thin films internal volume, often having a thickness from units to tens and hundreds of nanometres [7]. To adjust mechanical properties of thin films and membranes based on them, a new technology for controlling material grains orientation is used [8]. There are known works on formation of thin-film silicon membranes with subsequent study of mechanical properties of Si [9].
In this case, in order to improve reliability and performance of products, combination of layers is used, the overall structure of which consists of successively alternating thin metal-dielectric coatings [10]. However, such a multilayer cluster is affected by residual mechanical stresses, which are distributed both at the film-film and film-substrate interface and in thin-film structures thickness [11]. One of the ways to solve this problem is pretreatment of the surface with an ion beam.
Thin-film coatings with low surface roughness (units and fractions of nanometres) are widely used in optics, photonics and nanoelectronics. In particular, the fields of laser instrumentation and gyroscopy and, in particular, for the production of transistors on flexible substrates. It was shown in [12, 13] that it is possible to form layers with roughness parameters Sq less than 2.5 Å and Ra less than 1.0 nm by vacuum deposition methods.
The aim of this work is to evaluate influence of thin film coating thickness, substrate shape and topology on the residual stress level.
RESIDUAL STRESS MEASUREMENT TECHNIQUE
Residual mechanical stresses are a multifactorial property of thin film coatings. To calculate residual stresses, the Stoney formulae [14] are often used for small film thicknesses without the influence of external process factors (1):
, (1)
here σf – mechanical stress value, Pa;
Es – Young’s modulus of the plate material, Pa;
ds – plate thickness, m;
vs – Poisson’s ratio of the plate material;
df – layer thickness, m;
Rп, Rд – radius of curvature of the plate surface after and before, m, and [15] for models for coating deposition with substrate heating (2):
, (2)
here σf – mechanical stress value, Pa;
Es – Young’s modulus of the plate material, Pa;
αf, αs – temperature coefficient of expansion of the coating and the plate, 1/°C ;
ds – plate thickness, m;
∆T– deposition temperature, °C;
vs – Poisson’s ratio of the plate material;
df – layer thickness, m.
The analysis of the formulas shows possibility of obtaining empirical values of mechanical stresses by measuring the curvature radius of the plate, as well as obtaining data by means of analytical calculation on the thermal stresses influence.
There is a variety of factors for appearance of mechanical stresses in thin films, which depend on the film structures thickness [16], on the temperature regimes of formation [17], on the methods of formation (magnetron sputtering [18, 19], atomic layer deposition by ion assisted deposition [20]).
The main disadvantage of presence of stresses is relaxation state, which results in formation of defects in the thin film structure. Such defects include: cracking and segmentation (tensile stresses), delamination and warping (both tensile and compressive stresses), corrugation (compressive stresses), and deformation relief (both tensile and compressive stresses) [21].
The following reasons for residual stresses formation in structures are identified: mismatch between the lattice and monoepitaxial film layers (due to relaxation of the structure); surface tension of island grains (appearance of compressive stresses); coalescence of islands (appearance of tensile stresses); diffusion mobility of adatoms; appearance and annihilation of defects (due to vacancies, dislocations); phase transformations (due to doping, introduction of impurities) [22, 23].
Figure 1 shows peculiarity of stress formation in a thin film on a thick silicon substrate. Surface effects introduce compressive (negative) mechanical stresses at the substrate/film boundary σгр. Positive mechanical stresses σобъем are then generated in the film volume, with the total stress in the film σпленки being the sum of σгр and σобъем. Since the substrate/film system is at rest, σпленки = – σпластины.
PROCESS EQUIPMENT
Formation of thin film coatings is carried out on the MVTU-11-1MS unit (Fig.2) located at the Department of Electronic Technologies in Mechanical Engineering of Bauman Moscow State Technical University [24].
A cylindrical metal chamber with a volume of 22 litres and a two-stage pumping system allows cleaning and preparation of substrates with subsequent formation of functional layers in a single vacuum cycle, which ensures defect-free quality of structures.
The magnetron system with a target size of 50.8 mm operates from a power supply with an automatic matching device in the high-frequency plasma mode (13.56 MHz). The ion-beam processing system consists of a cylindrical ion source with a cold cathode, with a beam diameter of 23 mm, and a power supply operating in current stabilisation mode.
The machine fulfils the requirements of modern laboratories: oil-free high vacuum, flexible control system, short idle and pumping times.
ANALYTICAL EQUIPMENT
As measuring equipment we use optical profilometer WYKO NT9300 (Fig.3), designed for non-contact measurement of surface roughness parameters, linear topography along x, y and z axes on the substrate diameter up to 200 mm with a relative error of ±2% with height differences up to 10 mm and three-dimensional visualisation of solid objects surface with reflection coefficient more than 1% by optical method based on interference of light beams reflected from the mirror and from the sample under study. The ultimate vertical resolution of the profilometer is 0.1 nm.
The studies were carried out in VSI (Vertical Scanning Interferometry) mode in white light.
EXPERIMENTAL PROCEDURE
The technological route of technological and analytical equipment operation is shown in Fig.4.
The technological cycle of the structure forming starts with preliminary liquid cleaning of the substrate in an ultrasonic bath. Piezoelements that provide excitation of the ultrasonic wave operate at a frequency of 120 kHz. The bath is also equipped with the possibility of heating the cleaning solution, controlling and maintaining the set parameters. Preliminary chemical degreasing is carried out in ammonia solution H2O2:NH4OH:H2O (2:1:10) by cavitation of the ultrasonic bath at 40 °C for 4 minutes. To dehydrate the substrate surface, it was cleaned in isopropyl alcohol of 99.8% purity for 2 minutes.
Final cleaning and preparation of substrates for deposition of structures takes place in a vacuum chamber pumped to a pressure of 9.0 · 10–3 Pa. Then the substrate is treated with a high-energy ion beam of the working gas (argon) at an accelerating voltage of up to 3000 V and a discharge current of 30 mA at a vacuum pressure of 1.1 · 10-1 Pa.
A coating of all-composite Al2O3 target is formed by magnetron sputtering method at a voltage frequency of 13.56 MHz and 60 W glow discharge power. The thickness of the Al2O3 film is evaluated and controlled by cross-sectional chipping of the witness sample on a scanning electron microscope.
RESEARCH METHODOLOGY
The algorithm [25], which is based on Stoney’s formula (1), is used as a method to study and estimate the magnitude of mechanical stresses by calculating the curvature radius of the substrate before and after coating formation.
The 76 mm diameter monocrystalline silicon substrate is measured with an optical profilometer midway along the base slice of the substrate and across before (Fig.5a) and after deposition of the Al2O3 coating (Fig.5b).
The measurements generate a pattern of substrate topography from multiple points (Fig.6a), which is converted to substrate curvature (Fig.6b). Similar measurements take place after vacuum deposition of the film. For the adequacy of the calculation, the introduced substrate curvature (Fig.6c) is taken into account, which is recalculated into the average values along the two measurement directions of the stress distribution across the substrate (Fig.6d).
The uniqueness of this calculation method lies in calculation of mechanical stresses in the local region of the plate.
These studies qualitatively evaluate the shape of the substrates used. It is revealed that polished and prepared substrates have convex and concave shapes, which should also be taken into account in the development of technology for thin film coatings deposition.
RESULTS AND DISCUSSION
As a result of experimental study cycle, the dependence of the mechanical stress in the film on the Al2O3 layer thickness is formed (Fig.7).
Single-crystalline silicon substrates (Fig.7a–d) with thickness of 370±10 μm were used for experimental study. The analysis of the profilograms allowed us to divide the substrates into groups according to their shape: convex and concave. It is noticeable that with increasing film thickness, the degree of influence of the initial shape of the silicon substrate surface decreases.
Negative values of stress values indicate compression stresses. The data correlate with [26, 27], which show the results of measuring mechanical stresses during dielectric structures formation.
Analysis of the measurement results shows that surface effects introduce compressive (negative) mechanical stresses at the substrate/film boundary. Then positive mechanical stresses are formed in the film volume. This character of dependence coincides with the trends obtained in other works for thin film aluminium and molybdenum [8].
The analyses of the thin film structure using preliminary ion cleaning revealed that the roughness (Fig.8) of silicon decreases from 12.6±0.1 to 9.5±0.1 nm and roughness of Si-Al2O3 structure decreases from 17.8±0.2 to 13.9±0.2 nm.
CONCLUSIONS
Evaluation of mechanical stresses allows to reveal some key features, regularities and to form recommendations on the technology of vacuum deposition of Al2O3 thin film coating. As a result of the work, the method of surface roughness minimisation by means of surface pretreatment with a high-energy ion beam has been tested.
Distribution of mechanical stresses in a thin film of aluminium oxide over a silicon wafer is presented. By examining the optical profilometer surface, it was found that the roughness value decreases for the ion-cleaned structure. As the film thickness increases, the degree of influence of the initial surface shape of the silicon wafer decreases.
ACKNOWLEDGEMENTS
The work was financially supported by the Russian Federation represented by the Ministry of Science and Higher Education (agreement No. 075-15-2021-1350 dated 5 October 2021, internal number 15.SIN.21.0004).
PEER REVIEW INFO
Editorial board thanks the anonymous reviewer(s) for their contribution to the peer review of this work. It is also grateful for their consent to publish papers on the journal’s website and SEL eLibrary eLIBRARY.RU.
Declaration of Competing Interest. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Modern technological solutions in various areas of mechanical engineering, robotics, medicine and other industrial fields cannot be imagined without rapidly developing field of nanoengineering, which includes the thin-film coating formation technology.
Different types of thin films are distinguished according to their functional purpose and areas of use. For example, conductive layers are formed in printed circuit boards and power electronics for load transfer and switching between electrical parts [1, 2]. Insulating layers serve to form a barrier area beyond which no electrical power propagation takes place [3]. Thin films that enhance protective, antifriction, lubrication and corrosion-resistant properties of parts are used for deposition on the working surfaces of cutting tools, roll surfaces and equipment that is operated in aggressive oxygen-containing environments [4]. Optical thin film coatings – anti-reflective, fluorescent, reflective, etc., have found their application in solving everyday and design problems [5]. Also, decorative structures are actively used to create unique patterns and inscriptions [6].
It should be noted that practical implementation of nanoscale thin film coatings, which occurs due to the dimensional effect – the effect of enhancing the level of physical parameters of the material (electrical conductivity, magnetic, optical and mechanical properties) due to predominance of the surface area over the thin films internal volume, often having a thickness from units to tens and hundreds of nanometres [7]. To adjust mechanical properties of thin films and membranes based on them, a new technology for controlling material grains orientation is used [8]. There are known works on formation of thin-film silicon membranes with subsequent study of mechanical properties of Si [9].
In this case, in order to improve reliability and performance of products, combination of layers is used, the overall structure of which consists of successively alternating thin metal-dielectric coatings [10]. However, such a multilayer cluster is affected by residual mechanical stresses, which are distributed both at the film-film and film-substrate interface and in thin-film structures thickness [11]. One of the ways to solve this problem is pretreatment of the surface with an ion beam.
Thin-film coatings with low surface roughness (units and fractions of nanometres) are widely used in optics, photonics and nanoelectronics. In particular, the fields of laser instrumentation and gyroscopy and, in particular, for the production of transistors on flexible substrates. It was shown in [12, 13] that it is possible to form layers with roughness parameters Sq less than 2.5 Å and Ra less than 1.0 nm by vacuum deposition methods.
The aim of this work is to evaluate influence of thin film coating thickness, substrate shape and topology on the residual stress level.
RESIDUAL STRESS MEASUREMENT TECHNIQUE
Residual mechanical stresses are a multifactorial property of thin film coatings. To calculate residual stresses, the Stoney formulae [14] are often used for small film thicknesses without the influence of external process factors (1):
, (1)
here σf – mechanical stress value, Pa;
Es – Young’s modulus of the plate material, Pa;
ds – plate thickness, m;
vs – Poisson’s ratio of the plate material;
df – layer thickness, m;
Rп, Rд – radius of curvature of the plate surface after and before, m, and [15] for models for coating deposition with substrate heating (2):
, (2)
here σf – mechanical stress value, Pa;
Es – Young’s modulus of the plate material, Pa;
αf, αs – temperature coefficient of expansion of the coating and the plate, 1/°C ;
ds – plate thickness, m;
∆T– deposition temperature, °C;
vs – Poisson’s ratio of the plate material;
df – layer thickness, m.
The analysis of the formulas shows possibility of obtaining empirical values of mechanical stresses by measuring the curvature radius of the plate, as well as obtaining data by means of analytical calculation on the thermal stresses influence.
There is a variety of factors for appearance of mechanical stresses in thin films, which depend on the film structures thickness [16], on the temperature regimes of formation [17], on the methods of formation (magnetron sputtering [18, 19], atomic layer deposition by ion assisted deposition [20]).
The main disadvantage of presence of stresses is relaxation state, which results in formation of defects in the thin film structure. Such defects include: cracking and segmentation (tensile stresses), delamination and warping (both tensile and compressive stresses), corrugation (compressive stresses), and deformation relief (both tensile and compressive stresses) [21].
The following reasons for residual stresses formation in structures are identified: mismatch between the lattice and monoepitaxial film layers (due to relaxation of the structure); surface tension of island grains (appearance of compressive stresses); coalescence of islands (appearance of tensile stresses); diffusion mobility of adatoms; appearance and annihilation of defects (due to vacancies, dislocations); phase transformations (due to doping, introduction of impurities) [22, 23].
Figure 1 shows peculiarity of stress formation in a thin film on a thick silicon substrate. Surface effects introduce compressive (negative) mechanical stresses at the substrate/film boundary σгр. Positive mechanical stresses σобъем are then generated in the film volume, with the total stress in the film σпленки being the sum of σгр and σобъем. Since the substrate/film system is at rest, σпленки = – σпластины.
PROCESS EQUIPMENT
Formation of thin film coatings is carried out on the MVTU-11-1MS unit (Fig.2) located at the Department of Electronic Technologies in Mechanical Engineering of Bauman Moscow State Technical University [24].
A cylindrical metal chamber with a volume of 22 litres and a two-stage pumping system allows cleaning and preparation of substrates with subsequent formation of functional layers in a single vacuum cycle, which ensures defect-free quality of structures.
The magnetron system with a target size of 50.8 mm operates from a power supply with an automatic matching device in the high-frequency plasma mode (13.56 MHz). The ion-beam processing system consists of a cylindrical ion source with a cold cathode, with a beam diameter of 23 mm, and a power supply operating in current stabilisation mode.
The machine fulfils the requirements of modern laboratories: oil-free high vacuum, flexible control system, short idle and pumping times.
ANALYTICAL EQUIPMENT
As measuring equipment we use optical profilometer WYKO NT9300 (Fig.3), designed for non-contact measurement of surface roughness parameters, linear topography along x, y and z axes on the substrate diameter up to 200 mm with a relative error of ±2% with height differences up to 10 mm and three-dimensional visualisation of solid objects surface with reflection coefficient more than 1% by optical method based on interference of light beams reflected from the mirror and from the sample under study. The ultimate vertical resolution of the profilometer is 0.1 nm.
The studies were carried out in VSI (Vertical Scanning Interferometry) mode in white light.
EXPERIMENTAL PROCEDURE
The technological route of technological and analytical equipment operation is shown in Fig.4.
The technological cycle of the structure forming starts with preliminary liquid cleaning of the substrate in an ultrasonic bath. Piezoelements that provide excitation of the ultrasonic wave operate at a frequency of 120 kHz. The bath is also equipped with the possibility of heating the cleaning solution, controlling and maintaining the set parameters. Preliminary chemical degreasing is carried out in ammonia solution H2O2:NH4OH:H2O (2:1:10) by cavitation of the ultrasonic bath at 40 °C for 4 minutes. To dehydrate the substrate surface, it was cleaned in isopropyl alcohol of 99.8% purity for 2 minutes.
Final cleaning and preparation of substrates for deposition of structures takes place in a vacuum chamber pumped to a pressure of 9.0 · 10–3 Pa. Then the substrate is treated with a high-energy ion beam of the working gas (argon) at an accelerating voltage of up to 3000 V and a discharge current of 30 mA at a vacuum pressure of 1.1 · 10-1 Pa.
A coating of all-composite Al2O3 target is formed by magnetron sputtering method at a voltage frequency of 13.56 MHz and 60 W glow discharge power. The thickness of the Al2O3 film is evaluated and controlled by cross-sectional chipping of the witness sample on a scanning electron microscope.
RESEARCH METHODOLOGY
The algorithm [25], which is based on Stoney’s formula (1), is used as a method to study and estimate the magnitude of mechanical stresses by calculating the curvature radius of the substrate before and after coating formation.
The 76 mm diameter monocrystalline silicon substrate is measured with an optical profilometer midway along the base slice of the substrate and across before (Fig.5a) and after deposition of the Al2O3 coating (Fig.5b).
The measurements generate a pattern of substrate topography from multiple points (Fig.6a), which is converted to substrate curvature (Fig.6b). Similar measurements take place after vacuum deposition of the film. For the adequacy of the calculation, the introduced substrate curvature (Fig.6c) is taken into account, which is recalculated into the average values along the two measurement directions of the stress distribution across the substrate (Fig.6d).
The uniqueness of this calculation method lies in calculation of mechanical stresses in the local region of the plate.
These studies qualitatively evaluate the shape of the substrates used. It is revealed that polished and prepared substrates have convex and concave shapes, which should also be taken into account in the development of technology for thin film coatings deposition.
RESULTS AND DISCUSSION
As a result of experimental study cycle, the dependence of the mechanical stress in the film on the Al2O3 layer thickness is formed (Fig.7).
Single-crystalline silicon substrates (Fig.7a–d) with thickness of 370±10 μm were used for experimental study. The analysis of the profilograms allowed us to divide the substrates into groups according to their shape: convex and concave. It is noticeable that with increasing film thickness, the degree of influence of the initial shape of the silicon substrate surface decreases.
Negative values of stress values indicate compression stresses. The data correlate with [26, 27], which show the results of measuring mechanical stresses during dielectric structures formation.
Analysis of the measurement results shows that surface effects introduce compressive (negative) mechanical stresses at the substrate/film boundary. Then positive mechanical stresses are formed in the film volume. This character of dependence coincides with the trends obtained in other works for thin film aluminium and molybdenum [8].
The analyses of the thin film structure using preliminary ion cleaning revealed that the roughness (Fig.8) of silicon decreases from 12.6±0.1 to 9.5±0.1 nm and roughness of Si-Al2O3 structure decreases from 17.8±0.2 to 13.9±0.2 nm.
CONCLUSIONS
Evaluation of mechanical stresses allows to reveal some key features, regularities and to form recommendations on the technology of vacuum deposition of Al2O3 thin film coating. As a result of the work, the method of surface roughness minimisation by means of surface pretreatment with a high-energy ion beam has been tested.
Distribution of mechanical stresses in a thin film of aluminium oxide over a silicon wafer is presented. By examining the optical profilometer surface, it was found that the roughness value decreases for the ion-cleaned structure. As the film thickness increases, the degree of influence of the initial surface shape of the silicon wafer decreases.
ACKNOWLEDGEMENTS
The work was financially supported by the Russian Federation represented by the Ministry of Science and Higher Education (agreement No. 075-15-2021-1350 dated 5 October 2021, internal number 15.SIN.21.0004).
PEER REVIEW INFO
Editorial board thanks the anonymous reviewer(s) for their contribution to the peer review of this work. It is also grateful for their consent to publish papers on the journal’s website and SEL eLibrary eLIBRARY.RU.
Declaration of Competing Interest. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Readers feedback
rus


