Properties of reactively sputtered W–B–N thin film as a diffusion barrier for Cu metallization on Si

Thin films of W–B–N (10 nm) have been evaluated as diffusion barriers for Cu interconnects. The amorphous W–B–N thin films were prepared at room temperature via reactive magnetron sputtering using a W₂B target at various N₂/(Ar + N₂) flow ratios. Cu diffusion tests were performed after in-situ deposition of 200 nm Cu. Thermal annealing of the barrier stacks was carried out in vacuum at elevated temperatures for one hour. X-ray diffraction patterns, sheet resistance measurement, cross-section transmission electron microscopy images, and energy-dispersive spectrometer scans on the samples annealed at 500°C revealed no Cu diffusion through the barrier. The results indicate that amorphous W–B–N is a promising low resistivity diffusion barrier material for copper interconnects.     

Differential AC-chip calorimeter for glass transition measurements in ultra thin polymeric films

A differential AC-chip calorimeter capable to measure the glass transition in nanometer thin films is described. Due to the differential setup pJ/K sensitivity is achieved. Heat capacity can be measured for sample masses below one nanogram even above room temperature as needed for the study of the glass transition in nanometer thin polymeric films. The calorimeter allows for the frequency dependent measurement of complex heat capacity in the frequency range from 1 Hz to 1 kHz. The glass transition in thin films of polystyrene (PS) (100–4 nm) and polymethylmethacrylate (PMMA) (400–10 nm) was determined at well defined experimental time scales. No thickness dependency of the glass transition temperature was observed within the error limits (±3 K) - neither at constant frequency nor for the traces in the activation diagrams (1 Hz–1 kHz).     

Pulsed electron-beam melting of a copper — steel 316 system: Evolution of the chemical composition,microstructure, and properties

The surface topography, chemical composition, microstructure, nanohardness, and tribological characteristics of a Cu (film, 512 nm)-stainless steel 316 (substrate) system subjected to pulsed melting by a low-energy (20–30 keV), high-current electron beam (2–3 μs, 2–10 J/cm²) were investigated. The film was deposited by sputtering a Cu target in the plasma of a microwave discharge in argon. To prevent local exfoliation of the film due to cratering, the substrate was multiply pre-irradiated with 8–10 J/cm². On single irradiation, the bulk of the film survived, and a diffusion layer containing the film and substrate components was formed at the interface. The thickness of this layer was 120–170 nm irrespective of the energy density. The diffusion layer consisted of subgrains of γ-Fe solid solution and nanosized particles of copper. In the surface layer of thickness 0.5–1 μm, which included the copper film quenched from melt and the diffusion layer, the nanohardness and the wear resistance nonmonotonicly varied with energy density, reaching, respectively, a maximum and a minimum in the range 4.3–6.3 J/cm². As the number of pulsed melting cycles was increased to five in the same energy density range, there occurred mixing of the film-substrate system and a surface layer of thickness ∼2 μm was formed which contained ∼20 at. % copper. Displacement of the excess copper during crystallization resulted in the formation of two-phase nanocrystal interlayers separating the γ-phase grains. __________ Translated from Izvestiya Vysshikh Uchebnykh Zavedenii, Fizika, No. 12, pp. 6–13, December, 2005.     

The effect of the averaged structural and energetic features on the cohesive energy of nanocrystals

The size dependency of the cohesive energy of nanocrystals is obtained in terms of their averaged structural and energetic properties, which are in direct proportion with their cohesive energies. The significance of the effect of the geometrical shape of nanoparticles on their thermal stability has been discussed. The model has been found to have good prediction for the case of Cu and Al nanoparticles, with sizes in the ranges of 1–22 nm and 2–22 nm, respectively. Defining a new parameter, named as the surface-to-volume energy-contribution ratio, the relative thermal stabilities of different nanoclusters and their different surface-crystalline faces are discussed and compared to the molecular dynamic (MD) simulation results of copper nanoclusters. Finally, based on the size dependency of the cohesive energy, a formula for the size-dependent diffusion coefficient has been presented which includes the structural and energetic effects. Using this formula, the faster-than-expected interdiffusion/alloying of Au_(core)–Ag_(shell) nanoparticles with the core–shell structure, the Au-core diameter of 20 nm and the Ag-shell thickness of 2.91 nm, has been discussed and the calculated diffusion coefficient has been found to be consistent with its corresponding experimental value.     

Carbon nanotubes as nanoparticles collector

This communication reports on a new method for the collection of nanoparticles using carbon nanotubes (CNT) as collecting surfaces, by which the problem of agglomeration of nanoparticles can be circumvented. CNT (10–50 nm in diameter, 1–10 μm in length) were grown by thermal CVD at 923 K in a 7 v/v% C₂H₂ in N₂ mixture on electroless nickel-plated copper transmission electron microscopy (TEM) grids and Monel coupons. These samples were then placed downstream of an arc plasma reactor to collect individual copper nanoparticles (5–30 nm in diameter). It was observed that the Cu nanoparticles preferentially adhere onto CNT and that the macro-particles (diameter >1 μm), a usual co-product obtained with metal nanoparticles in the arc plasma synthesis, are not collected. Cu–Ni nanoparticles, a catalyst for CNT growth, were deposited on CNT to grow multibranched CNT. CNT-embedded thin films were produced by re-melting the deposited nanoparticles.     

Molecular-dynamics simulation of the heat capacity for nickel and copper clusters: Shape and size effects

We have investigated the heat capacity of ideal Cu and Ni fcc clusters with diameters from 2 to 6 nm in the temperature range 200–800 K by the molecular-dynamics method using a modified tight-binding potential. Our analysis has shown consistency with the experimental results at temperatures of 200–300 K. The data obtained are also indicative of several regularities that are in agreement with the analytical calculations. We have concluded from the results of our computer simulations that the heat capacity in the case of isolated free clusters can exceed that of a bulk material, with this difference decreasing as the nanoparticle grows proportionally to the reduction in the fraction of surface atoms. The excess of the heat capacity for ideal copper and nickel nanoclusters with D = 6 nm at T = 200 K has been found to be 10% and 13%, respectively. Consequently, the large heat capacities of copper and nickel nanostructures observed in some real experiments cannot be related to the characteristics of free clusters. We hypothesize that these properties of a nanomaterial depend on the degree of agglomeration of its constituent particles, i.e., the surfaces and interphase boundaries of interconnected nanoclusters can have a strong effect. To test this hypothesis, we took nickel and copper clusters of various sizes (4000–7200 atoms) produced through the simulation of condensation from the gas phase. At high temperatures, we failed to adequately assess the role of the interphase boundaries in calculating the heat capacity of nanoparticles. The reason was the mass diffusion of Ni or Cu atoms to impart an energetically more favorable shape and structure to the synthesized clusters. At low temperatures, the heat capacity of such clusters exceeded that of clusters with an ideal shape and structure by a value from 3.2% to 10.6%. We have concluded that the Ni and Cu clusters produced in real experiments cannot be applied in devices using the thermal energy of such clusters without a preliminary optimization stage, because their external shape and interior structure are nonideal.     

Optical and IR studies on r.f. magnetron sputtered ultra-thin MoO3 films

Ultra-thin MoO₃ films were deposited onto glass and Si substrates by r.f. magnetron sputtering. The optical and IR properties of the films were studied in the range of 250 to 1000 nm and 400 to 1500 cm⁻¹, respectively. The optical transmission spectra show a significant shift in absorption edge. The energy gap of the films deposited at 373 K and 0.1 mbar was found to be 3.93 eV, and it decreases with increasing substrate temperature and decreasing sputtering pressure. The IR transmittance spectra shows strong modes of vibrations of Mo=O and Mo–O–Mo units of MoO₃ molecule. A significant change in energy gap and a shift in frequency of IR modes were observed in ultra-thin MoO₃ films.     

Atom beam sputtered Mo2C films as a diffusion barrier for copper metallization

The saddle field fast atom beam sputtered (ABS) 50 nm thick molybdenum carbide (Mo₂C) films as a diffusion barrier for copper metallization were investigated. To study the diffusion barrier properties of Mo₂C films, the as-deposited and annealed samples were characterized using four probes, X-ray diffraction, field enhanced scanning electron microscopy, energy dispersive X-ray analysis, atomic force microscopy and Rutherford back scattering techniques. The amorphous structure of the barrier films along with presence of carbon atoms at the molybdenum carbide-silicon interface is understood to reduce effective grain boundaries and responsible for increased thermal stability of Cu/Mo₂C/Si structure. The lowest resistivity of the as-deposited molybdenum carbide barrier films was ∼29 μΩ cm. The low carbon containing molybdenum carbide was found thermally stable up to 700 °C, therefore can potentially be used as a diffusion barrier for copper metallization.     

Influence of thermal diffusion on the laser ablation of thin polymer films

The laser ablation of a photosensitive triazene polymer was investigated with a ns XeCl excimer laser over a broad range of thicknesses (10–400 nm). We found that the ablation threshold fluence increased dramatically with decreasing film thickness for films thinner than 50 nm. Ablation on substrates with different thermal properties (sapphire, fused silica, PMMA) was investigated as well, and a clear influence of the substrate material was obtained. A mathematical model combining thermal diffusion and absorption effects was used to explain the experimental data. The model is in good agreement with the experimental data and shows that heat diffusion into the substrate plays a crucial role for the ablation process of very thin films. PACS 52.38.Mf; 44.05.+e; 81.05.Lg     

Observation of slowly decreasing molecular oscillations in ultrathin liquid films using X-ray reflectivity

We studied ~0.5 μm and 30–80 ? thick films of a normal dielectric liquid, tetrakis(2-ethylhexoxy)silane (TEHOS), at temperature range 228–286 K, deposited onto silicon (111) substrate with native oxide using X-ray reflectivity. TEHOS is spherical with size ~10 ?, non-polar, non-reactive, and non-entangling; TEHOS has been reported to show interfacial layering at room temperature and surface layering at 0.23 T_c (T_c≈ 950 K). For films m thick, the reflectivity data did not change significantly as a function of temperature; for films 30–80 ? thick, the reflectivity data did change. The data could be fitted with an electron density model composed of a minimum necessary number of Gaussians and a uniform density layer with error-function broadened interfaces. When the film thickness is 60–80 ? below 246 K, we found that the interface and the surface layering coexist but do not overlap. When the film thickness is 30–40 ? below 277 K, they overlap and the electron density profile shows slowly decreasing molecular oscillations at the air-liquid interface.     

Spectral studies of the structure of copper phthalocyanin-polyimide composites produced by vacuum deposition

Visible and IR absorption spectra are studied for copper phthalocyanin (CuPc)-polyyimide composites. The composite films are obtained by vacuum codeposition of CuPc, pyromellitedianhydride, and diaminodiphenyloxide followed by temperature-induced imidization. On the basis of IR spectra it is shown that in the prepared composites imidization occurs over the whole CuPc concentration range (10–90%). From the visible spectra it is found that CuPc is deposited mainly in the form of a-microcrystals, which are destroyed during imidization to produce noncrystalline molecular aggregates. Subsequent heat treatment (320–350°C) gives rise to CuPc β-microscrystals. A. N. Sevchenko Research Institute of Applied Physical Problems, 7, Kurchatov St., 220064, Minsk, Belarus. Translated from Zhurnal Prikladnoi Spektroskopii, Vol. 64, No. 4, pp. 502–506, July–August, 1997.     

Adsorption of titanium,chromium, and copper atoms on thin aluminum and magnesium oxide film surfaces

Methods of Auger electron spectroscopy (AES), spectroscopy of characteristic electron energy losses (SCEEL), slow electron diffraction (SED), and contact potential difference (CPD) in ultrahigh vacuum are used to investigate the adsorption-emission properties and stability of two-component film systems formed by putting of Ti, Cr, and Cu atoms on MgO–Mo(011) and Al₂O₃–Mo(011) surfaces. All atoms have the properties of electronegative adsorbates. Continuous adatom monolayers are formed on the Al₂O₃–Mo(011) system surface, and three-dimensional islands are formed on the MgO–Mo(011) surface. The properties of monoatomic films on the oxide layer surface are close to those observed for bulk materials. No radical changes of the system properties are detected with increasing dielectric layer thickness. The thermal stability of the newly formed structures decreases in the order Ti, Cr, Cu, Al₂O₃(MgO), and Mo(011).     

Photoluminescence of ZnS:Cu crystals at various concentrations of copper

Doping of ZnS crystals with background impurities and ZnS:Al crystals by various concentrations of copper from a bismuth melt has been carried out. The photoluminescence spectra of the starting ZnS crystals annealed in the bismuth melt and doped by copper have been investigated. Interpretation of the experimental results according to the model in which the associates (Cu _Zn ^′ Cu _i ^● ) are responsible for the B-Cu band (≈460 nm) and the donor—receptor pairs [Cu _i ^● -(Cu _Zn ^′ Cu _i ^● )] are responsible for the G-Cu band (≈505 nm) is given. It is assumed that doping by copper from a bismuth melt is accompanied by the separation of a Cu_xS-type phase. It has been shown that heat treatment of ZnS crystals in the Bi melt does not lead to the appearance of luminescence centers based on Bi_Zn, Bi_S, and Bi_i. __________ Translated from Zhurnal Prikladnoi Spektroskopii, Vol. 72, No. 6, pp. 794–798, November–December, 2005.     

LaF<Subscript>3</Subscript>:Er,Yb doped sol–gel polymeric optical waveguide amplifiers

Oleic acid modified LaF₃:Er,Yb nanocrystals have been synthesized with a molar ratio of La:Yb:Er=75:20:5 and dispersed in toluene with 40 wt. % sol–gel derived organic–inorganic hybrid materials for films. The nanocomposite films showed strong 1540-nm luminescence intensity under the excitation of 980 nm after heat treatment below 150 °C. The full width at half maximum bandwidth was 51-nm wide around 1540 nm. We also fabricated reverse-mesa ridge waveguides using LaF₃:Er,Yb nanocrystals containing sol–gel materials. With an input pump of 150 mW, a signal enhancement of 2.3 dB at the wavelength of 1550 nm was measured in a 1.3-cm-long waveguide. PACS 42.00.00; 42.70.-a; 42.82.Et     

Nickel oxide thin films synthesized by reactive pulsed laser deposition: characterization and application to hydrogen sensing

Transparent nickel oxide thin films were grown by reactive pulsed laser deposition. An ArF* (λ=193 nm, τ=12 ns) excimer laser source was used to ablate the Ni targets in a controlled pressure of ambient oxygen. The substrates were either kept at room temperature or heated to a selected temperature within the 200–400 °C range. Post-deposition heat treatment, which was applied to further promote crystallization and overcome any oxygen deficiency, yielded transparent thin films. The surface morphology and crystalline status of the synthesized thin structures were analyzed in correlation with their optical properties. A significant response to several concentrations of hydrogen was demonstrated when heating the nickel oxide films at 185 °C. PACS 78.66.Hf; 81.15.Fg; 82.47.Rs     

New method for determining the thermal diffusion coefficient of thin films using ferroelectric crystals

A new method for determining the thermal diffusion coefficient of thin dielectric films deposited on a ferroelectric crystal detector is presented. The surface of the specimen was periodically heated by a rectangular modulated heat flow. The impact of the ratio of the film thickness to the detector thickness, as well as the film thermal diffusion coefficient on the shape of the system’s pyroelectric response is analyzed. Optimal conditions for the measurements are offered. Approbation of the method was carried out on an unannealed lead zirconate titanate (PZT) film deposited on a lithium tantalate crystal. The thermal diffusion coefficient of the specimen was α = 1.5 × 10⁻⁷ m²/s, which correlates with an independent estimation of this magnitude for polarized films.     

Core-shell LaF<Subscript>3</Subscript>:Er,Yb nanocrystal doped sol–gel materials as waveguide amplifiers

The oleic acid (OA)-modified LaF₃:Er,Yb and LaF₃:Er,Yb–LaF₃ core-shell nanocrystals are synthesized. The lifetime values could be further improved by incorporating core-shell nanocrystals. A kind of sol–gel derived organic–inorganic hybrid material (SGHM) allows for 50 wt.% or even more of both the two nanocrystals in the matrix, and we give the explanation from scattering analysis. It’s precisely because we use the erbium nanocrystals rather than erbium organic complexes, and avoid the undesirable luminescence quenching by Er–Er clustering with a high Er³⁺ concentration. LaF₃:Er,Yb–LaF₃/SGHM transparent films and optical waveguides are also fabricated. The nanocomposite films show strong 1550 nm luminescence intensity under the excitation of 980 nm after heat treatment below 150 °C and the full-width-at-half-maximum is about 51 nm. The loss and optical gain of the waveguide are measured. A relative gain of about 3.5 dB is measured at 1550 nm in a 1.7 cm long waveguide.     

The structure of copper-doped amorphous hydrogenated carbon films

A study of the nanostructure of a-C:H:Cu films by x-ray small-angle scattering, x-ray diffraction, TEM, and visible and UV spectroscopy is reported. It has been established that introduction of 9–16 at.% Cu not only decorates the original carbon fragments but produces extended (up to 4 μm in length) formations of copper-decorated strongly elongated ellipses as well. At 14–16 at.% Cu, these linear clusters represent copper nanotubes with a core made up of the original ellipses drawn in a line. It is these conducting copper formations that account primarily for the strong increase in conductivity at 12–16 at.% Cu contents in a-C:H films. Fiz. Tverd. Tela (St. Petersburg) 41, 2088–2096 (November 1999)     

Segmental dynamics of poly(methyl phenyl siloxane) confined to nanoporous glasses

The effect of a nanometer confinement on the molecular dynamics of poly(methyl phenyl siloxane) (PMPS) was studied by dielectric spectroscopy (DS), temperature modulated DSC (TMDSC) and neutron scattering (NS). Nanoporous glasses with pore sizes of 2.5–20 nm have been used. DS and TMDSC experiments show that for PMPS in 7.5 nm pores the molecular dynamics is faster than in the bulk which originates from an inherent length scale of the underlying molecular motions. For high temperatures the temperature dependence of the relaxation rates for confined PMPS crosses that of the bulk state. Besides finite states effects also the thermodynamic state of nano-confined PMPS is different from that of the bulk. At a pore size of 5 nm the temperature dependence of the relaxation times changes from a Vogel/Fulcher/Tammann like to an Arrhenius behavior where the activation energy depends on pore size. This is in agreement with the results obtained by NS. The increment of the specific heat capacity at the glass transition depends strongly on pore size and vanishes at a finite length scale between 3 and 5 nm which can be regarded as minimal length scale for glass transition to appear in PMPS.     

Theoretical model for describing degradation of thin hydrogen-containing films

A new theoretical model is proposed to describe the behavior of films of composite hydrogen-containing compounds during heat treatment. The model is based on the thermal generation of hydrogen atoms and atoms of the composite compounds with their subsequent diffusion to the boundaries of the film and percolation through the surface into the atmosphere. Calculations are performed for the heat treatment of films of silicon nitride. A comparison with the experimental data in the literature demonstrates the high efficiency of the proposed model. Zh. Tekh. Fiz. 67, 105–110 (August 1997)