Thermal effects during the deposition of thin silver,gold and copper films and their influence on internal stress measurements

The temperature changes of glass substrates were measured during the deposition of silver, gold and copper films. The measurements were carried out continuously and in situ throughout the evaporation process. In addition, the absorptivities of metal films of various thicknesses were determined for black-body radiation at 300 and 1300 K after deposition. The results that were obtained are used to calculate the variation of the substrate temperature during the evaporation of the metal. This in turn permits the correction of mechanical stress measurements for inherent thermal effects.     

The influence of O2, H2, H2O,N2, CO and CH4 on the structure of thin silver films investigated by ultrahigh vacuum internal stress measurements

It has been demonstrated earlier that the structure of thin metal films can be investigated continuously during deposition by measuring the internal film stress. With these experiments it has been shown that various deposition parameters (evaporation rate, ambient atmosphere etc.) considerably influence the internal stress and thus the structure of the evaporated metal films. In order to separate the various parameters determining film growth we developed an ultrahigh vacuum version of the stress measuring apparatus used previously. An application of this method to silver films deposited onto fresh MgF₂ substrate films in various gas ambients is described. The film structure deduced from the stress curves on the basis of a model for the origin of the internal stress is compared with the structure visible in the electron microscope.     

Microstructural evolution in nanostructured gold films

Thin Au films in the thickness range t = 1.5-126 nm were coated by DC sputtering on SiO_x/Si substrates at room temperature inside a vacuum chamber with a base pressure of about 1 × 10^− 2 mbar (1 Pa). The film structure, nanograin characteristics, and the surface roughness as a function of thickness were analyzed using X-ray diffraction, scanning tunneling microscopy and transmission electron microscopy. The results reflect the microstructural evolution with film thickness. They help us to understand the mode of grain growth, which is monomodal-normal columnar as well as spherical. By determination of the dynamic scaling exponent derived from power law dependence of the mean grain size and film thickness, the prevailed mechanism of grain growth is deduced to be the diffusion of mobile Au atoms in grain boundaries. The surface roughness increases with the film thickness following a power law: R_rms ~ t^b. The linear fitted value for b is 0.60.     

Electron microscope structure and internal stress in thin silver and gold films deposited onto MgF2 and SiO substrates

In this work the internal stress of thin silver and gold films deposited onto MgF₂ and SiO substrate films was measured up to a mean thickness of about 1000 Å. The method used was based on the cantilever beam principle. The internal stress of the metal films was determined by subtracting the contributions due to momentum transfer and heat transfer. From parallel electron microscope investigations of films of various thicknesses it was seen that a close correlation exists between the internal stress generated in the films and their structure. On the basis of these results we propose a model which gives a qualitative explanation of the measured internal stress versus thickness curves and of their dependence on vacuum conditions, evaporation rate and the nature of the substrate.     

Thin,quench-condensed lead-based alloy films investigated by resistivity and superconducting tunneling measurements

Quench-condensed lead-based alloy films were investigated by the superconductive tunneling method to study how structural order affects the phonon distribution, and in particular the appearance of localized impurity modes, As solute, 10 at. % of either Na, Mg, Zn, Ga, Ge, Ag, Cd, In, Sn, Sb, or Te was used. The films were condensed upon a substrate at 0.4 K and studied as deposited and after anneals to 30, 100, and 300 K. It was found that the shape of the phonon-induced tunneling structure could be used as a means to classify the state of structural order in the films. Temperatures of transformations from one state to another could be identified with those where steps occurred in curves of film resistance vs. annealing temperature. The most notable transformation was the one that took the film from an amorphous to a finely grained crystalline state. The transformation temperature scales with the melting point of the pure impurity element of the binary alloy. As the annealing temperature was increased further another temperature region was reached where the crystallites grew in size, causing notable changes in the phonon spectra. A third transformation occurred at low temperatures within the amorphous state for a few of the alloys.Localized impurity modes could be observed only inPbIn and only when the films had been annealed so that relatively large crystallites had formed. The absence of these high-frequency modes in most of the alloy films can be explained as being due to a lack of order or solubility in the ordered state. Such arguments cannot, however, explain why no localized modes are seen inPbSn.In disordered films the phonon distribution is shifted toward lower frequencies compared to the case of a single crystal. For this type of lead-based alloy the shift implies lower values of the superconducting transition temperatureT _c and gap ₀ (0), but a higher2 ₀ (0O)/k_B T _C ratio. The dependence of these values and of the tunnel resistance upon the structural state is studied. The effect of fluctuations on the superconducting transition is noted for the thin, amorphous films.     

Growth and structure of epitaxial films of rh deposited on sodium chloride,silver, copper and gold

Single crystal thin films of rhodium are prepared by thermal evaporation of the metal on (100) air cleaved NaCl faces or on epitaxial films of Cu, Ag or Au. (100) rhodium //(100) substrate orientations are always dominant. The epitaxial temperature is lower and better monocrystalline deposits are formed on the metallic substrate     

Melting and crystallization of copper, silver, and gold droplets

The crystallization kinetics in 0.5-g droplets of melted copper, silver, and gold (special purity grade metals) cooled in vacuum at a rate of 0.01 K/s was studied. Under these experimental conditions, the physical supercooling observed in all three metals prior to crystallization was virtually zero. An analysis of the values of supercooling obtained in this study together with the data obtained previously under the same conditions showed that supercooling increases in a regular manner with the number of electrons on the outermost p level.     

Internal stress of thin silver and gold films and its dependence on gas adsorption

Using the highly sensitive bending beam device described in earlier papers the internal stress versus thickness curves of silver and gold have been measured continuously during metal deposition. The contributions due to momentum transfer and heat transfer have been subtracted from the experimentally determined curves. Using the model for the origin of the internal stress which we have developed previously, the measured dependence of the stress curves on the nature of the substrate, on the evaporation rate for the metal deposition and on gas adsorption by the substrate, as well as by the metal itself, are interpreted in terms of changes in the structure of the metal film.     

Thermal expansion of copper,silver, and gold at low temperatures

Improvements have been made in a differential dilatometer using the three-terminal capacitance detector. The dilatometer is of copper and has been calibrated from 1.5–34 K in an extended series of observations using silicon and lithium fluoride as low-expansion reference materials. The expansion of silver and gold samples has been measured relative to the dilatometer, while the calibrations themselves have been used to determine the expansion of copper relative to the reference materials. Analyses of six sets of observations indicate that below 12 K the linear expansion coefficient α of copper is represented by $$10^{10} \alpha = (2.1_5 \pm 0.1){\rm T} + (0.284 \mp 0.005){\rm T}^3 + (5 \pm 3) \times 10^{ - 5} T^5 K^{ - 1} $$ corresponding to respective electronic and lattice Grüneisen parameters γ _e =0.9 ₃ and γ ₀ =γ ₁ =1.78. Measurements on oxygen-free silver yield $$10^{10} \alpha = (1.9 \pm 0.2){\rm T} + (1.14 \mp 0.03){\rm T}^3 + (2 \pm 2) \times 10^{ - 4} T^5 K^{ - 1} $$ below 7 K, whence γ _e ? 0.9 ₇ , γ ₀ =γ ₁ =2.23. By contrast, silver containing ca. 0.02 at. % oxygen showed a much larger expansion at the lowest temperatures: below 7 K, 10 ¹⁰ α ～ 7T+1.19T ³ . We have not been able to obtain an unambiguous representation for gold, but find a reasonable fit below 7 K to be $$10^{10} \alpha \simeq (1 \pm 0.5){\rm T} + (2.44 \mp 0.05){\rm T}^3 - (5 \pm 1) \times 10^{ - 3} T^5 K^{ - 1} $$ with γ ₁ ? 2.94 and γ _e ? 0.7 (free-electron value).     

Thermal expansion of copper,silver, and gold below 30 K

A variable transformer technique has been used to determine the linear thermal expansion coefficients of the noble metals from 4 to 30 K. The precision of the data initially was ±0.04 Å, and this was later increased to ±0.015 Å, resulting in a sensitivity of approximately 2×10^?11 for relative length changes of a 10-cm-long sample. The results agree at all temperatures (to better than 5%) with those of White and Collins who used a differential capacitor technique. The differences are 2% or less above 20 K for Cu, above 8 K for Ag, and at all temperatures for Au. The differences between the two sets of data for the three metals are not systematic (α_WC greater for Cu, less for Ag) and may be due to differences in sample purity since much larger low-temperature anomalies were found within each set for certain samples of Cu and Ag. The resulting electronic Grüneisen parameters γ_e and theT=0 lattice Grüneisen parameters γ₀ are as follows: $$\begin{gathered} copper \gamma _{\text{e}} = 0.91 \pm 0.05 \gamma _{\text{0}} = 1.67 \pm 0.02 \hfill \\ silver \gamma _{\text{e}} = 1.18 \pm 0.15 \gamma _{\text{0}} = 2.29 \pm 0.03 \hfill \\ gold \gamma _{\text{e}} = 1.6 \pm 0.5 \gamma _{\text{0}} = 2.96 \pm 0.04 \hfill \\ \end{gathered} $$ The values of γ₀ are in reasonable (5% at worst) agreement with elastic constant values.     

Computer simulation of surface diffusion of copper, silver and gold

The binding energies to copper, silver and gold (111) surfaces of self-atom clusters have been calculated. The activation energies of motion of these ad-atom clusters, vacancies and divacancies on copper, silver and gold (111) surface, and of the conversion of ad-atom clusters on (111) and (100) have been calculated by use ofn-body embedded atom potentials and molecular dynamics.     

Adsorption of thin isobutane films on silicon investigated by X-ray reflectivity measurements

An X-ray reflectivity study of the silicon isobutane interface is presented. Thin isobutane films were adsorbed on a silicon wafer and their film thickness was measured as a function of pressure in order to determine the effective Hamaker constant which is a proportion of the coupling between substrate and adsorbed film. A comparison with theoretical expressions of the effective Hamaker constant is given.     

Resistance changes in discontinuous gold films

The change in the electrical resistance of vacuum-deposited island films of gold after deposition has been investigated. This change is found to be well explained by the morphological change of the island particles due to surface self-diffusion, provided that the electron transport between the island particles obeys the relationship derived by Neugebauer and Webb and that the temperature rise induced by the deposition decays exponentially with time.     

Ultra-microhardness of vacuum-deposited films II: Results for silver,gold,copper,MgF2, LiF and ZnS

The penetration depths ε of a diamond triangular cone indenter under loads W of 1–100 mgf have been studied for evaporated films and a bulk specimen of each material, making use of the ultra-microhardness tester described in Part I (Nishibori and Kinosita, Thin Solid Films, 48 (1978) 325). The ε versus W curves are featured by the presence of a critical load W_c below which ε is unobservable. W_c varies with the material and specimen preparation: it is generally of the order of 10 mgf but is smaller by at least one order for the metal films. When W exceeds W_c, ε increases as k₁^-1(W - W_c)^{$\text{1}\text{2}$} for film and bulk specimens of the metals and as k₂^-1(W - W_c) for those of the dielectrics. The constants k₁ and k₂, which can be regarded as measures of hardness, vary with the material and the specimen preparation.     

Ageing studies of discontinuous copper and silver thin films

The results of experiments carried out on the post-deposition resistance changes in discontinuous films of copper and silver with and without overlayers of SiO and Al₂O₃ are presented. The changes in the sheet resistance of the films with time and pressure were studied for the above combinations. Mobility coalescence is assumed to be responsible for the resistance increase of an uncovered copper film of initial resistance 1.9 M/. On exposure to the atmosphere, it was found that an Ag/SiO combination of initial resistance of 0.1 M/ achieved stability in the sheet resistance much quicker than a Cu/Al₂O₃ combination of initial resistance 20 M/. The fall in resistance of the Cu/Al₂O₃ composite is attributed to the formation of Al₂(OH)₆ due to the interaction of Al₂O₃ with the water vapour in atmosphere. Copper films with and without overlayers of Al₂O₃ show an abrupt increase in the sheet resistance as a function of pressure at a pressure of about 5 × 10^–2 torr with the maximum rate of change of resistance occurring at higher pressure for the higher resistance film. This indicates that the overlayer of Al₂O₃ is very porous in nature. Field effect studies were carried out on an uncovered copper film of initial resistance 10 M/ and the behaviour was found to be ohmic up to a field of 800 V cm^–1.     

Electrochemical detection of chloride by underpotentially deposited silver films on polycrystalline gold

This paper describes an electrochemical method for measuring dilute levels of chloride using an underpotentially deposited (UPD) Ag adlayer on polycrystalline Au substrates as a sensing agent. Specifically, chloride ions adsorb onto the Ag UPD adlayer and effect changes in the electrochemical deposition and stripping characteristics of the silver film. Cyclic voltammograms (CVs) of the native Au/Ag(UPD) electrode in 0.1 M H2SO4(aq) exhibit a primary stripping peak for the Ag UPD adlayer at 550 mV vs Ag(+/0), and chloride adsorption onto the Au/Ag(UPD) surface effects a peak shift to approximately 600 mV vs Ag(+/0), depending on the amount of adsorbed Cl-, as affected by the Cl- concentrations and contact times employed in the derivatization. The chloride-treated electrodes also exhibit a stripping peak at 275 mV that is not observed on the native substrate and increases in intensity with Cl- concentration and derivatization time. The integrated charge density for this latter stripping peak relative to that for the primary stripping peak at 550-610 mV provides a useful metric for quantifying adsorbed Cl- levels, and these values allow measurement of Cl- concentrations in dilute aqueous solutions. For Cl- concentrations between 0.5 and 100 microM, the kinetics of Cl- adsorption followed a transient Langmuir adsorption model and allowed measured surface coverages to be used for determining Cl- solution concentrations. Using contact times of 1 min for Cl- adsorption, the electrodes showed a linear response across Cl- concentrations of 0.5-20 microM.     

Structure and internal stress in ultra-thin silver films deposited on MgF2 and SiO substrates

A substantially improved bending-beam technique combined with the capacitance method was used for the continuous determination of the internal stress during vacuum deposition of ultra-thin silver films. With this method a sensitivity improvement of about two orders of magnitude was achieved compared with results reported previously. By proper experimental design all radiation interference on the stress-measuring appratus except the unavoidable direct heat radiated from the evaporators was eliminated. With this improved method it was possible to determine continuously the internal stress of silver films deposited onto fresh MgF₂ or SiO substrate films under varying experimental conditions (i.e. evaporation rate, residual gas pressure). Based on correlations between internal stress and the structure of the films observed using an electron microscope a model is proposed that explains the origin of the alternately compressive and tensile stress in such ultra-thin films.     

An investigation of silver diffusion through gold films by Auger electron spectroscopy

Silver diffusion through gold films of different thicknesses was observed by Auger electron spectroscopy at room temperature. It was found that very thin enriched silver layers have occurred on gold films within one month after the evaporation of gold films on silver substrates. Some experiments for determining the effects of sulphur adsorption layers on the thin films are also reported.     

Auger analysis of thick silver films and of silver layers grown on copper

A sensitive new measure of the line broadening of an Auger doublet, called the R factor, was applied to the growth of (1) thick silver films and (2) thin silver films on a thick copper substrate. It is shown that periodic fluctuations in R with thickness give evidence for monocrystalline layer growth while the absence of such fluctuations is characteristic of polycrystalline growth. By combining measurements of the Auger amplitude and the R factor as functions of the silver film thickness for silver films deposited onto copper, it was shown that a Stranski-Krastanov growth mode is characteristic of the epitaxial growth of Ag(111) on Cu(111) at 210°C. The initial layer in this case is two (111) atom spacings thick on which high flat islands form. Continued growth of these islands is shown to take place by two-atom-thick flat islands.