rus
Issue #3-4/2022
V.A.Kazakov, A.V.Smirnov, A.V.Kokshina, E.S.Tyunterov, V.S.Abrukov, D.A.Anufrieva
SYNTHESIS AND STUDY OF THE HYBRID METAL-CARBON SYSTEMS OPTICAL PROPERTIES: LINEAR-CHAIN CARBON FILMS DOPED WITH SILVER
SYNTHESIS AND STUDY OF THE HYBRID METAL-CARBON SYSTEMS OPTICAL PROPERTIES: LINEAR-CHAIN CARBON FILMS DOPED WITH SILVER
DOI: 10.22184/1993-8578.2022.15.3-4.186.194
The paper describes the synthesis technology and the results of studying optical properties of metal-carbon systems: films of linear-chain carbon doped with silver with optical spectrophotometry and spectral ellipsometry. The results of modeling and generalization of the obtained data with the help of artificial neural networks are presented.
The paper describes the synthesis technology and the results of studying optical properties of metal-carbon systems: films of linear-chain carbon doped with silver with optical spectrophotometry and spectral ellipsometry. The results of modeling and generalization of the obtained data with the help of artificial neural networks are presented.
Теги: artificial neural networks linear-chain carbon metallocarbon nanosystems silver искусственные нейронные сети линейно-цепочечный углерод металлуглеродные наносистемы серебро
Received: 20.05.2022 | Accepted: 27.05.2022 | DOI: https://doi.org/10.22184/1993-8578.2022.15.3-4.186.194
Original paper
SYNTHESIS AND STUDY OF THE HYBRID METAL-CARBON SYSTEMS OPTICAL PROPERTIES: LINEAR-CHAIN CARBON FILMS DOPED WITH SILVER
V.A.Kazakov1, Cand. of Sci. (Tech), Associate Professor, ORCID: 0000-0001-8974-2307 / cossac@mail.ru
A.V.Smirnov1, Engineer, ORCID: 0000-0003-2424-8142
A.V.Kokshina1, Senior Lecturer, ORCID: 0000-0001-8645-2822
E.S.Tyunterov1, Post Graduate, ORCID: 0000-0002-8816-9737
V.S.Abrukov1, Doctor of Sci. (Physics and Mathematics), Prof., ORCID: 0000-0002-4680-6224
D.A.Anufrieva1, Post-graduate, ORCID: 0000-0003-4860-3460
Abstract. The paper describes the synthesis technology and the results of studying optical properties of metal-carbon systems: films of linear-chain carbon doped with silver with optical spectrophotometry and spectral ellipsometry. The results of modeling and generalization of the obtained data with the help of artificial neural networks are presented.
Keywords: metallocarbon nanosystems, linear-chain carbon, silver, artificial neural networks
For citation: V.A. Kazakov, A.V. Smirnov, A.V. Kokshina, E.S. Tyunterov, V.S. Abrukov, D.A. Anufrieva. Synthesis and study of the hybrid metal-carbon systems optical properties: linear-chain carbon films doped with silver. NANOINDUSTRY. 2022. V. 15, no. 3–4. PP. 186–194. https://doi.org/10.22184/1993-8578.2022.15.3-4.186.194
INTRODUCTION
Synthesis of the allotropic form of carbon in sp1 state was firstly performed in INEOS RAS USSR in 1959 by oxidative dehydropolycondensation of acetylene [1, 2]. According to electron microdiffraction data [2] two modifications of carbin, named a- and b-carbin, were found. The hexagonal cell parameters were determined as aa= 5.08 Å, сa= 7.80 Å, аb= 4.76 Å, сb= 2.58 Å, respectively. The authors assumed that a-carbin is the polyyne form of the carbon chain ( -С≡С-С≡С- )n, and b-carbin is the cumulene form ( =С=С=С= )n. Continued research of both natural and synthetically produced carbins has significantly expanded the range of possible modifications of carbin, in particular, a stable carbin nanofilm with unique physical properties has been obtained. This film has been named linear-chain carbon (LCC) film, or sp1 hybridised carbon. An important area of further work was the development of technologies for introduction of metal atoms into the LCC film [3–6]. The study of structural and optical properties of nanomaterials obtained by the introduction of metal atoms into the LCC film is of great fundamental and applied importance, since the obtained hybrid nanosystems may be of interest in medicine [3, 4], catalysis and development of new nanosensors, nanoelectronics products [5–7]. Optical properties of nano-thick films of metals and their nanoparticles manifest themselves in appearance of a plasmon effect. For silver, this effect is characteristic in the visible wavelength range. When the carbon matrix is doped with silver or silver films are annealed as a result of diffusion processes and surface tension forces upon heating, the dimensions of nanoparticles and clusters change. In addition, during formation of silver films and clusters by different methods their size depends significantly on the synthesis parameters. For example, in thermo-resistive sputtering the film growth takes place according to the cluster mechanism: the first stage is the formation of clusters and silver nanoparticles on the substrate, which are the nucleation centres of the film. The film itself is not continuous and transparent, as it is composed of individual nanoparticles, and the "thickness" of the film is determined by the cluster size or, calculated equivalently, by a change in the film mass. If there is no overlap of the individual nanoparticles, its electrical resistance is high. As the thickness increases, the silver clusters increase in size, their overlap occurs and silver film becomes solid, which manifests itself in the appearance of electrical conductivity. The mechanism of plasmons formation changes (from particles to film), which is manifested in optical properties as well.
RESEARCH METHODS
The experiment on silver embedding into the film of linear-chain carbon (LCC) was constructed as follows. In UVR-3M vacuum apparatus at a residual pressure of about 10–2 Pa a 10÷100 nm thick silver film was deposited on glass substrates by thermo-resistive evaporation in vacuum. On the part of samples, a LCC film was synthesized on top of silver in the upgraded vacuum ion-plasma unit "URM.3.279.070 Almaz" using the method described in [3]. Thicknesses were measured by atomic force microscopy, by measuring a step at the glass/silver film interface. Thickness of the LCС was determined by the number of pulses of carbon plasma and equaled approximately 100 nm. The resulting bilayer sample was placed in a furnace and annealed in air at 450 °C. Annealing was carried out to stimulate the introduction of silver atoms into the LCС.
The transmission coefficient spectra of pure silver films and silver-LCС metal films obtained before and after annealing were studied with a Lambda-25 spectrophotometer.
The dielectric permittivity spectra of pure silver films (real and imaginary parts) were examined on an Ellips-1891 spectral ellipsometer.
Presence or absence of electrical conductivity of the films was confirmed on a Keythley 2400 series digital programmable multimeter.
RESULTS
For silver films sprayed on a glass substrate (or quartz glass), at different thicknesses of the sprayed silver layer, the optical transmission spectra are as shown in Fig.1.
For continuous (completely covering the substrate, low surface roughness), thin (up to 200 nm thickness), transparent, electrically conductive silver films, without annealing, the maximum transmittance is observed at a wavelength of 322 nm (Fig.1, red spectrum).
For islets (not completely covering the substrate, large surface roughness with the formation of individual silver clusters on the surface of the substrate), thin (up to 70 nm thick), transparent non-conductive silver films, annealed in air, besides a higher maximum transmittance at 322 nm, a wide extinction band with a centre at 468 nm is observed (Fig.1, green spectrum). After annealing, as a result of diffusion processes and surface tension forces, the film is rearranged, the continuity is broken, the film becomes islet-like and surface plasmonic absorption appears [13]. The centre of the absorption band determines the size of silver nanoparticles [9].
For the same silver islet film (green spectrum) as in Fig.1, Fig.2 shows the spectra of the real and imaginary parts of the dielectric permittivity of the film: eps1 and eps2 (by definition, the dielectric permittivity eps = eps1 + i·eps2).
As can be seen from Fig.2, starting with an energy of 3.8 eV (322 nm) there is a step change both in the dielectric constant and in the correspondingly dependent values of the refractive index and extinction. This frequency corresponds to the plasmonic frequency, i.e., the light with a frequency below the plasma frequency is reflected because electrons in the metal shield the electric field in a the light electromagnetic wave. The light with a frequency above the plasma frequency passes through because the electrons cannot respond fast enough to shield it. It is the bulk plasmons that explain the transparency window (Fig.3) with a maximum at a wavelength of 322 nm [9].
Figure 4 shows transmission spectra for 60 nm thick silver film systems with the applied 100 nm LCC film. The grey line corresponds to the system before annealing and the black line to the system after annealing in air.
As can be seen from Fig.4, in the silver-LCC system after annealing there were significant changes in the transmittance spectrum. In addition to an increase in transmittance over almost the entire wavelength range and a maximum of transparency at 322 nm, a pronounced absorption band centered at ~420 nm appeared.
Thus, the obtained experimental results show the following:
Thus, during thermal annealing of a silver film with LCC there is a rearrangement of silver islet-like film accompanied by an increase in the size of silver clusters, which is reflected in the transmission spectra. This process is accompanied by introduction of clusters and nanoparticles into the LCC structure. It is known that direct interaction of carbon with silver does not form a chemical compound between these elements. The electronogram obtained in [10] and atomic force microscopy data [11] suggest that despite the changes in the surface structure of the film associated with the substance transfer, the hexagonal structure of LCC has not been destroyed during intercalation, as six reflexes characteristic of LCC films are observed.
DISCUSSIONS
In order to summarize all the obtained experimental data (not only those described above), a set of multifactor computational models has been created using artificial neural networks (ANNs). The basics of ANNs and methods of using ANNs for simulating the experimental data have been presented and described in [14–17]. Neural networks included in the Deductor Academic 5.3 Build 0.88 analytics platform (www.basegroup.ru) were used to create the models.
We used feed forward neural networks with one input layer (the number of neurons was determined by the number of factors), one hidden layer with different number of hidden neurons (from 5 to 8), and a single output layer (with one neuron). The target functions of the models were: transparency coefficient and imaginary and real parts of dielectric constant. The factors of the models were: a light wavelength, surface structure of the silver film (solid or insular), thickness of the silver film, thickness of the LCC film, electrical conductivity (yes, no) and annealing of the films (yes, no). The obtained models were peculiar (obtained in the process of neural networks training) calculators (computational structures), allowing to determine the target function of a particular model for any set of factors values.
In addition to the calculator model the developed computational models determine the real and imaginary parts of the dielectric constant of different silver films through film parameters, as well as a computational model showing the relation between transparency coefficient and dielectric constant (real and imaginary parts) for different films produced by different synthesis techniques were also created.
Thus, the obtained computational neural network models correspond to the obtained experimental data. The value of the RMS error is determined with the aid of the analytical platform Deductor Academic in the process of ANN training, simultaneously for the training and test sample of the experimental data [18]. Calculation results for the most different combinations of factor values are in good agreement with the experimental results.
CONCLUSIONS
Detailed optical studies of silver films and silver-LCC hybrid metal-carbon systems have been carried out. The obtained results enabled to highlight the peculiarities of transmission and dielectric permittivity spectra. Strong influence of annealing on the spectra has been demonstrated. The artificial neural networks helped to develop the multifactorial computational models of the obtained spectra.
The authors see the use of ANNs in modelling experimental data and creating multivariate computational models as a very promising way to generalize the experimental results.
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.
Original paper
SYNTHESIS AND STUDY OF THE HYBRID METAL-CARBON SYSTEMS OPTICAL PROPERTIES: LINEAR-CHAIN CARBON FILMS DOPED WITH SILVER
V.A.Kazakov1, Cand. of Sci. (Tech), Associate Professor, ORCID: 0000-0001-8974-2307 / cossac@mail.ru
A.V.Smirnov1, Engineer, ORCID: 0000-0003-2424-8142
A.V.Kokshina1, Senior Lecturer, ORCID: 0000-0001-8645-2822
E.S.Tyunterov1, Post Graduate, ORCID: 0000-0002-8816-9737
V.S.Abrukov1, Doctor of Sci. (Physics and Mathematics), Prof., ORCID: 0000-0002-4680-6224
D.A.Anufrieva1, Post-graduate, ORCID: 0000-0003-4860-3460
Abstract. The paper describes the synthesis technology and the results of studying optical properties of metal-carbon systems: films of linear-chain carbon doped with silver with optical spectrophotometry and spectral ellipsometry. The results of modeling and generalization of the obtained data with the help of artificial neural networks are presented.
Keywords: metallocarbon nanosystems, linear-chain carbon, silver, artificial neural networks
For citation: V.A. Kazakov, A.V. Smirnov, A.V. Kokshina, E.S. Tyunterov, V.S. Abrukov, D.A. Anufrieva. Synthesis and study of the hybrid metal-carbon systems optical properties: linear-chain carbon films doped with silver. NANOINDUSTRY. 2022. V. 15, no. 3–4. PP. 186–194. https://doi.org/10.22184/1993-8578.2022.15.3-4.186.194
INTRODUCTION
Synthesis of the allotropic form of carbon in sp1 state was firstly performed in INEOS RAS USSR in 1959 by oxidative dehydropolycondensation of acetylene [1, 2]. According to electron microdiffraction data [2] two modifications of carbin, named a- and b-carbin, were found. The hexagonal cell parameters were determined as aa= 5.08 Å, сa= 7.80 Å, аb= 4.76 Å, сb= 2.58 Å, respectively. The authors assumed that a-carbin is the polyyne form of the carbon chain ( -С≡С-С≡С- )n, and b-carbin is the cumulene form ( =С=С=С= )n. Continued research of both natural and synthetically produced carbins has significantly expanded the range of possible modifications of carbin, in particular, a stable carbin nanofilm with unique physical properties has been obtained. This film has been named linear-chain carbon (LCC) film, or sp1 hybridised carbon. An important area of further work was the development of technologies for introduction of metal atoms into the LCC film [3–6]. The study of structural and optical properties of nanomaterials obtained by the introduction of metal atoms into the LCC film is of great fundamental and applied importance, since the obtained hybrid nanosystems may be of interest in medicine [3, 4], catalysis and development of new nanosensors, nanoelectronics products [5–7]. Optical properties of nano-thick films of metals and their nanoparticles manifest themselves in appearance of a plasmon effect. For silver, this effect is characteristic in the visible wavelength range. When the carbon matrix is doped with silver or silver films are annealed as a result of diffusion processes and surface tension forces upon heating, the dimensions of nanoparticles and clusters change. In addition, during formation of silver films and clusters by different methods their size depends significantly on the synthesis parameters. For example, in thermo-resistive sputtering the film growth takes place according to the cluster mechanism: the first stage is the formation of clusters and silver nanoparticles on the substrate, which are the nucleation centres of the film. The film itself is not continuous and transparent, as it is composed of individual nanoparticles, and the "thickness" of the film is determined by the cluster size or, calculated equivalently, by a change in the film mass. If there is no overlap of the individual nanoparticles, its electrical resistance is high. As the thickness increases, the silver clusters increase in size, their overlap occurs and silver film becomes solid, which manifests itself in the appearance of electrical conductivity. The mechanism of plasmons formation changes (from particles to film), which is manifested in optical properties as well.
RESEARCH METHODS
The experiment on silver embedding into the film of linear-chain carbon (LCC) was constructed as follows. In UVR-3M vacuum apparatus at a residual pressure of about 10–2 Pa a 10÷100 nm thick silver film was deposited on glass substrates by thermo-resistive evaporation in vacuum. On the part of samples, a LCC film was synthesized on top of silver in the upgraded vacuum ion-plasma unit "URM.3.279.070 Almaz" using the method described in [3]. Thicknesses were measured by atomic force microscopy, by measuring a step at the glass/silver film interface. Thickness of the LCС was determined by the number of pulses of carbon plasma and equaled approximately 100 nm. The resulting bilayer sample was placed in a furnace and annealed in air at 450 °C. Annealing was carried out to stimulate the introduction of silver atoms into the LCС.
The transmission coefficient spectra of pure silver films and silver-LCС metal films obtained before and after annealing were studied with a Lambda-25 spectrophotometer.
The dielectric permittivity spectra of pure silver films (real and imaginary parts) were examined on an Ellips-1891 spectral ellipsometer.
Presence or absence of electrical conductivity of the films was confirmed on a Keythley 2400 series digital programmable multimeter.
RESULTS
For silver films sprayed on a glass substrate (or quartz glass), at different thicknesses of the sprayed silver layer, the optical transmission spectra are as shown in Fig.1.
For continuous (completely covering the substrate, low surface roughness), thin (up to 200 nm thickness), transparent, electrically conductive silver films, without annealing, the maximum transmittance is observed at a wavelength of 322 nm (Fig.1, red spectrum).
For islets (not completely covering the substrate, large surface roughness with the formation of individual silver clusters on the surface of the substrate), thin (up to 70 nm thick), transparent non-conductive silver films, annealed in air, besides a higher maximum transmittance at 322 nm, a wide extinction band with a centre at 468 nm is observed (Fig.1, green spectrum). After annealing, as a result of diffusion processes and surface tension forces, the film is rearranged, the continuity is broken, the film becomes islet-like and surface plasmonic absorption appears [13]. The centre of the absorption band determines the size of silver nanoparticles [9].
For the same silver islet film (green spectrum) as in Fig.1, Fig.2 shows the spectra of the real and imaginary parts of the dielectric permittivity of the film: eps1 and eps2 (by definition, the dielectric permittivity eps = eps1 + i·eps2).
As can be seen from Fig.2, starting with an energy of 3.8 eV (322 nm) there is a step change both in the dielectric constant and in the correspondingly dependent values of the refractive index and extinction. This frequency corresponds to the plasmonic frequency, i.e., the light with a frequency below the plasma frequency is reflected because electrons in the metal shield the electric field in a the light electromagnetic wave. The light with a frequency above the plasma frequency passes through because the electrons cannot respond fast enough to shield it. It is the bulk plasmons that explain the transparency window (Fig.3) with a maximum at a wavelength of 322 nm [9].
Figure 4 shows transmission spectra for 60 nm thick silver film systems with the applied 100 nm LCC film. The grey line corresponds to the system before annealing and the black line to the system after annealing in air.
As can be seen from Fig.4, in the silver-LCC system after annealing there were significant changes in the transmittance spectrum. In addition to an increase in transmittance over almost the entire wavelength range and a maximum of transparency at 322 nm, a pronounced absorption band centered at ~420 nm appeared.
Thus, the obtained experimental results show the following:
- Silver nanofilms have a maximum transmittance at 322 nm;
- Annealing of silver films and silver-LCC systems leads to an increase of transmittance in the spectral region centered at 322 nm;
- Annealing of silver films and silver-LCC systems leads to appearance of absorption band centered at ~420 nm, which is typical for surface plasmonic absorption of silver.
Thus, during thermal annealing of a silver film with LCC there is a rearrangement of silver islet-like film accompanied by an increase in the size of silver clusters, which is reflected in the transmission spectra. This process is accompanied by introduction of clusters and nanoparticles into the LCC structure. It is known that direct interaction of carbon with silver does not form a chemical compound between these elements. The electronogram obtained in [10] and atomic force microscopy data [11] suggest that despite the changes in the surface structure of the film associated with the substance transfer, the hexagonal structure of LCC has not been destroyed during intercalation, as six reflexes characteristic of LCC films are observed.
DISCUSSIONS
In order to summarize all the obtained experimental data (not only those described above), a set of multifactor computational models has been created using artificial neural networks (ANNs). The basics of ANNs and methods of using ANNs for simulating the experimental data have been presented and described in [14–17]. Neural networks included in the Deductor Academic 5.3 Build 0.88 analytics platform (www.basegroup.ru) were used to create the models.
We used feed forward neural networks with one input layer (the number of neurons was determined by the number of factors), one hidden layer with different number of hidden neurons (from 5 to 8), and a single output layer (with one neuron). The target functions of the models were: transparency coefficient and imaginary and real parts of dielectric constant. The factors of the models were: a light wavelength, surface structure of the silver film (solid or insular), thickness of the silver film, thickness of the LCC film, electrical conductivity (yes, no) and annealing of the films (yes, no). The obtained models were peculiar (obtained in the process of neural networks training) calculators (computational structures), allowing to determine the target function of a particular model for any set of factors values.
In addition to the calculator model the developed computational models determine the real and imaginary parts of the dielectric constant of different silver films through film parameters, as well as a computational model showing the relation between transparency coefficient and dielectric constant (real and imaginary parts) for different films produced by different synthesis techniques were also created.
Thus, the obtained computational neural network models correspond to the obtained experimental data. The value of the RMS error is determined with the aid of the analytical platform Deductor Academic in the process of ANN training, simultaneously for the training and test sample of the experimental data [18]. Calculation results for the most different combinations of factor values are in good agreement with the experimental results.
CONCLUSIONS
Detailed optical studies of silver films and silver-LCC hybrid metal-carbon systems have been carried out. The obtained results enabled to highlight the peculiarities of transmission and dielectric permittivity spectra. Strong influence of annealing on the spectra has been demonstrated. The artificial neural networks helped to develop the multifactorial computational models of the obtained spectra.
The authors see the use of ANNs in modelling experimental data and creating multivariate computational models as a very promising way to generalize the experimental results.
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
