Search

In this paper, we report on the in situ synthesis of graphene layers by means of chemical vapor deposition (CVD), directly on nickel micro-electromechanical systems (MEMS) surfaces. We have developed MEMS structures of which the temperature can be increased locally by Joule heating while in a methane environment. For our MEMS structures, the thermal time constant is 28 μs. As a result, we have control over the carbon precipitation time, thereby governing how many graphene layers are formed. Bi-layer to multi-layer graphene was observed using micro-Raman spectroscopy, but not single-layer graphene, as it gives no Raman signal when coupled on a nickel surface. The corresponding precipitation control theory is also presented in this paper, in which we relate the out-diffusion of carbon atoms from the grains of the nickel structure to the resulting number of graphene layers. Our method provides regulated carbon segregation from nickel and allows a prescribed number of graphene layers to form by tuning the precipitation time. In this way, we enable the direct in situ synthesis of graphene locally on the top and sidewalls of nickel MEMS structures, so that e.g. such graphene-coated MEMS surfaces can contribute towards a promising solution against friction and wear for MEMS devices with sliding components.

Single-crystal copper films on sapphire have recently been reported upon in relation to graphene growth on these films. In the present paper the kinetics of the formation of single crystal copper films is investigated. We demonstrate the importance of heating the sapphire substrate in 1000 hPa oxygen, followed by a fast cooling prior to depositing the copper film. The importance of this treatment is tentatively explained by the dissolution of oxygen in sapphire and subsequent out-diffusion during recrystallization of the copper film to form a copper-oxide interface layer. Also, the importance of avoiding oxygen incorporation in the sputter deposited film is demonstrated.

In this paper, we report on the influence of capillary condensation on the sliding friction of sidewall surfaces in polycrystalline silicon micro-electromechanical systems (MEMS). We developed a polycrystalline silicon MEMS tribometer, which is a microscale test device with two components subject to sliding contact. One of the components can be heated in situ by Joule heating to set the temperature of the contact and thereby control the capillary kinetics at the MEMS sidewalls. We used an optical displacement measurement technique to record the stick–slip motion of the slider with sub-nanometer resolution, and we assessed the friction force with nanonewton resolution. All friction measurements were performed under controlled ambient conditions while sweeping the contact temperature from room temperature to 300 °C, and from 300 °C to room temperature. We were able to distinguish the two ways in which energy is dissipated during sliding: the ‘semi-statically’ dissipated energy attributed to asperity deformation and contact yield, and the dynamically dissipated energy ascribed to the release of the tension in the slider during slip events. We observed an increase in the dynamically dissipated energy at 80 °C while sweeping down in temperature. This increase is caused by higher adhesion due to capillary condensation between the conformal surfaces. Our study highlights how energy is dissipated during the sliding contact of MEMS sidewalls, and it is helpful in overcoming friction in multi-asperity systems.

Abrasive corrosive wear of materials is a phenomenon with great economical consequences. Here we present abrasive corrosive wear experiments performed on structural steel. We compare pin on disc experiments in artificial seawater, deionized water and ethanol. Counterintuitively the wear is observed to be smallest in seawater. We ascribe this result to the combined chemical modification of the steel surface due to local electrochemical interaction with the seawater and the mechanical interaction between the pin and the surface.

We report a nanoinfrared (IR) imaging study of the localized plasmon resonance modes of graphene nanoribbons (GNRs) using a scattering-type scanning near-field optical microscope (s-SNOM). By comparing the imaging data of GNRs that are aligned parallel and perpendicular to the in-plane component of the excitation laser field, we observed symmetric and asymmetric plasmonic interference fringes, respectively. Theoretical analysis indicates that the asymmetric fringes are formed due to the interplay between the localized surface plasmon resonance (SPR) mode excited by the GNRs and the propagative surface plasmon polariton (SPP) mode launched by the s-SNOM tip. With rigorous simulations, we reproduce the observed fringe patterns and address quantitatively the role of the s-SNOM tip on both the SPR and SPP modes. Furthermore, we have seen real-space signatures of both the dipole and higher-order SPR modes by varying the ribbon width.

The wear of pipelines, used in slurry transport, results in high costs for maintenance and replacement. The wear mechanism involves abrasion, corrosion, impact, and the interaction among them. In this work, we study the effect of impact on the wear mechanism and wear rate. Results show that when the effect of impact is small, the wear mechanism is dominated by electrochemically induced surface modification, which leads to a lower wear rate in a corrosive environment than in a non-corrosive environment. By contrast, when the effect of impact is large, the wear mechanism is drastically altered. In that regime plastic deformation is important. The influence of corrosion in the high impact regime can be neglected. Our findings show the importance of including impact effect in the distinction of wear of slurry pipes.

Uniaxial materials whose axial and tangential permittivities have opposite signs are referred to as indefinite or hyperbolic media. While hyperbolic responses are normally achieved with metamaterials, hexagonal boron nitride (hBN) naturally possesses this property due to the anisotropic phonons in the mid-infrared. Using scattering-type scanning near-field optical microscopy, we studied polaritonic phenomena in hBN. We performed infrared nano-imaging of highly confined and low-loss hyperbolic phonon polaritons in hBN. The polariton wavelength was shown to be governed by the hBN thickness according to a linear law persisting down to few atomic layers [1]. Additionally, we carried out the modification of hyperbolic response in meta-structures comprised of a mononlayer graphene deposited on hBN [2]. Electrostatic gating of the top graphene layer allows for the modification of wavelength and intensity of hyperbolic phonon polaritons in bulk hBN. The physics of the modification originates from the plasmon-phonon coupling in the hyperbolic medium. Furthermore, we demonstrated the “hyperlens” for subdiffractional focusing and imaging using a slab of hBN [3]. References [1] S. Dai et al., Science, 343, 1125 (2014). [2] S. Dai et al., Nature Nanotechnology, 10, 682 (2015). [3] S. Dai et al., Nature Communications, 6, 6963 (2015).

The aluminium alloy 7075-T6 is widely used in engineering. In some applications, like slurry transport, corrosion and abrasion occur simultaneously, resulting in early material failure. In the present work, we investigated the combined effect of corrosion and wear on the aluminium alloy 7075-T6. We performed two series of wear experiments to vary the conditions and severity of corrosion environment: chemically by using ethanol, deionized water and seawater, and electrochemically by applying various potentials using a potentiostat in seawater. Results show that, in seawater, the wear rate was higher than in deionized and ethanol; and in the potentiostat experiments, at the anodic potentials, the wear rates were higher than at the open circuit potential and the cathodic potentials. Seawater is the most corrosive one among the three liquids and the corrosion products can be easily removed. When applying anodic potentials, corrosion is accelerated, and the higher wear rate confirms that higher corrosion rate leads to higher wear rate due to the formation and removal of corrosion products during tribocorrosion.

We report on nano-infrared (IR) imaging studies of confined plasmon modes inside patterned graphene nanoribbons (GNRs) fabricated with high-quality chemical-vapor-deposited (CVD) graphene on Al2O3 substrates. The confined geometry of these ribbons leads to distinct mode patterns and strong field enhancement, both of which evolve systematically with the ribbon width. In addition, spectroscopic nano-imaging in mid-infrared 850-1450 cm-1 allowed us to evaluate the effect of the substrate phonons on the plasmon damping. Furthermore, we observed edge plasmons: peculiar one-dimensional modes propagating strictly along the edges of our patterned graphene nanostructures., Comment: 10 pages, 8 figures

We present results of molecular dynamics simulations of graphene on Cu surfaces. Interactions were modelled with the charge-optimized many-body potential, which gives a reasonable though not flawless description of the graphene–Cu system. The interaction between Cu and complete graphene sheets is characterized by an ‘averaged out’ interaction at a large bonding distance. Many bonding characteristics are indifferent to the details of how the Cu surface atoms are arranged, including the surface orientation and even if the surface is solid or molten. Graphene edges have a strong interaction with the Cu substrates. Systems were modelled at various temperatures, ranging from 0 K to the Cu melting temperature. At high temperature we find that the presence of graphene slightly stabilizes the Cu surface and retards surface melting. After cooling down to room temperature, the Cu substrate is 1.7% smaller than the graphene due to difference thermal expansion coefficients. This leads to the formation of wrinkles in graphene. Single wrinkles experience only small migration barriers and are quite mobile. When multiple wrinkles intersect, they form immobile knots that hinder further movement of the connected wrinkles. The elastic energy of the wrinkles and knots due to bending of the graphene is determined.

Aluminum Alloy 5083 is increasingly used in the ship building industry due to its excellent corrosion resistance and its favorable mechanical properties. A significant application of this alloy is in high speed ferries, where the influence of dynamic loads is of great importance as well. However, the influence of a corrosive environment and/or a welded condition on the fatigue properties of nautical aluminum alloys has been studied only superficially. In the present paper the influence of a seawater environment in combination with a welded state is described. Both fatigue limits and SN-curves are given, as well as a fracture surface analysis. From these results the fracture and crack initiation mechanisms are explained. Particularly the penetration of seawater into sub-surface pores, that are present in the welded material, results in a significant drop in fatigue life time when the welded alloy is tested in seawater.

Since long, the nitriding process has been used to improve the mechanical properties of the surface of steel. However, results from the prevailing model for the nitriding of steel are still not in perfect agreement with experimental data. It is investigated if this can be attributed to the model assumption that there is no interaction between nitrogen atoms and alloying atoms in solution in the iron matrix. To this end, the interaction of N with atomically dispersed Ti, V, Cr, Mo, and Ni alloying element atoms in ferritic steel is calculated using Density Functional Theory. Results of the Density Functional Theory calculations agree well with experimental lattice parameter, Mossbauer spectroscopy, and internal friction data taken from literature. Deficiencies in prevailing atomistic models for Fe-Me-N configurations in steel (Me denotes the alloying element considered) are revealed and the initial stages of MexN precipitation in alloyed iron are elucidated. For all elements but Ti, the interaction with N is negligible or repulsive. Significant attractive interaction exists only if Ti is a second nearest neighbour of N (−0.22 eV). It is shown that this interaction is too small to significantly influence nitriding at typical nitriding temperatures (> 500 °C ). Consequently, in modelling the nitriding of steel at typical nitriding temperatures, interaction between the atomically dispersed alloying elements and nitrogen can be disregarded. Internal friction experiments evidence nitrogen in solution in the steel bonded with higher binding energy, also for the other alloying elements. We suggest that this nitrogen is bound in Fe-Me-N accumulations consisting of more than one alloying atom and one or more nitrogen atoms. For the modelling of nitriding of steel, these interactions must be taken into account.

Forcefill is a method used in metallization of ICs to fill via holes with metal. At elevated temperature and under applied pressure the metal flows into the via hole. The method is applicable to aluminum as well as to copper. In this article the mechanism of the process is discussed based on measurements of the kinetics and on finite element calculations of the shear and hydrostatic stress in the film covering the via hole. It is demonstrated that naive use of deformation maps constructed by Frost and Ashby for the description of the forcefill process easily leads to highly inconsistent results. The calculated stress distribution shows that the relevant shear stress is more than one order of magnitude lower that the applied pressure. Since no deformation map was available for the grain size appropriate for the forcefill experiment a map for 1 μm grains was constructed. This map indicates that for the forcefill process diffusional flow is the dominant deformation mechanism. It is shown that under these conditions the diffusional flow process described by Frost and Ashby is in essence identical to our previously reported stress-induced diffusion model.

Storage tests have been performed to obtain information on the influence of the passivation material and its geometry on the mechanical reliability of Al–Cu lines. During storage tests, the stress in the lines is tensile and depends on the passivation material and passivation geometry. The passivation was either a SiO2 layer or a SiNx layer. In both cases the influence of a conformal and a planarised passivation geometry has been studied. Passivating the lines with a material with a higher stiffness such as SiNx will increase the stress void density in the lines. Moreover, a conformal passivation layer induces less stress voids in the metal than a planarised passivation, deposited by a Dep/Etch method. The number of stress voids saturates within 24 h at 200°C. However, the voids continue to grow during longer storage times. When the lines are passivated with the lower stiffness material SiO2, stress voids have not been observed after storage testing.

A newly designed bulge-testing apparatus for the mechanical testing of the tensile properties of free-standing thin films has been constructed and tested. With this instrument it is possible to measure the elastic modulus, tensile and compressive growth stress, and plain strain yield strength in thin films. The setup features a high strain resolution (4E-10), and a high stress resolution, e.g., >0.2 MPa for bulge heights larger than 10 μm. In our setup, thin films are stressed by a differential gas pressure across the film, while the deflection is measured by a scanning laser beam. The scanning laser beam measures the curvature of the bulge rather than the bulge height. This makes the setup insensitive to a possible initial nonflatness of the film. This also provides the possibility to measure the growth stress of films that were deposited under a compressive stress. We show the results of measurements of the plane strain modulus on thin tungsten films and on both flat and nonflat aluminum (alloy) film sa...

A high density plasma chemical vapour deposition (HDP CVD) system based on electron cyclotron resonance (ECR) plasma excitation for deposition of inter metal dielectric (IMD) is presented. With the system deposition of SiO 2 and SiOF has been performed. The influence of pressure, Ar content in the flow, total flow, bias voltage, microwave power on gap fill capability and growth rate has been investigated. A figure of merit, the product of gap filling capability and growth rate is defined. In addition measurements of the uniformity of the composition over the wafer of the deposited SiO 2 and SiOF layers were performed. The dielectric constant of the layers was measured on SiOF films with different composition. The stability of these SiOF films was also analysed. This was done by treating the films with moisture and measuring composition before and after this treatment.

The increasing use of closely spaced narrow interconnections and generally reduced feature sizes in modern IC technology has led to a requirement for anisotropic etching of high performance metallizations. Etching processes addressing this requirement should exhibit good uniformity and reproducibility, a good selectivity over underlying materials and the mask to ensure a good pattern fidelity. WGe based barriers have good prospects in sub-micron contact barrier technology. The purpose of the present paper is to investigate the reactive ion etching (RIE) of blanket deposited AlCu/WGe/sub x//Ti films on SiO/sub 2/ in a chlorine based plasma chemistry employing a Cl/sub 2/, SiCl/sub 4/, He mixture. Various aspects of the metal stack etching including profile control, etch rate and selectivity to oxide are discussed in detail.

High Density Plasma (HDP) reactors are currently introduced for the deposition of SiO/sub 2/ and SiOF in submicron gaps between aluminium conductors in advanced integrated circuits. The main feature of HDP reactors is the ability to fill high aspect ratio spaces between aluminium conductors. This feature is achieved by employing a simultaneous deposition and sputtering process.

Recently, high pressure via-fill has been developed as a new method to manufacture ULSI sub-micron diameter Al contacts with a high aspect ratio. In this paper the fabrication of sub-micron Al vias using silicon oil (polymethylsiloxane) at high pressures is investigated. The high pressure via-fill process is performed on sub-micron diameter contacts patterned in SiO/sub 2/ on Si(100) and covered with a hot-sputtered bridging AlCu(O.5%)Si(1%) film on a Ti/TiN barrier layer. During the process samples are heated to a temperature in the range of 250-400/spl deg/C and pressurized between 60 and 200 MPa using silicon oil as transducer. It is found that in this temperature and pressure range contact-fill is possible.

The mechanical behaviour of AlV(0.1 at%)Pd(0.1 at%) films during thermal cycling is studied. Apart from films line patterns have also been studied. Passivated and unpassivated line patterns covering the whole wafer were produced. The curvature in line length and line width directions of the various wafers is determined in situ, during thermal cycling between 50°C and 400°C. For unpassivated AlVPd line structures the curvature in the line width direction is smaller than in the line length direction. Furthermore, a linear dependence between curvature and temperature has been obtained. For wafers with passivated AlVPd lines the curvature in the line length and line width directions are of the same order of magnitude. Again the curvatures depend linearly on the temperature in the whole temperature range between 50°C and 400°C.

Stress measurements on 0.5 μm copper films (with optional passivation) on Si(100) have been performed during thermal cycling between 50 and 400°C. During annealing the film stress follows the thermal stress as calculated from the thermomechanical properties of copper, silicon dioxide and silicon. A deviation from linear behaviour is found which is caused by plastic deformation (strengthening and relaxation). After the first anneal the stress temperature dependence is very similar in both sample types indicating that the deformation processes are alike. The actual stress values in the individual Cu, SiO layers at 50°C are found and Annealing at constant temperature shows that in passivated and non-passivated Cu stacks stress relaxation occurs at any temperature (between 50 and 400°C). From stress measurements on individual SiO layers it is shown that relaxation takes place primarily in the CVD-Cu layer.

The kinetics of the hydrogen reduction of tungsten hexafluoride were studied. The growth rate was measured as a function of the process conditions, which were determined in situ by means of laser Raman scattering. The partial pressures were measured at a height of 17 mm above the center of a 100 mm wafer in a cold-wall single-wafer low-pressure chemical vapor deposition reactor. The hydrogen reduction reaction was on the order of 1/6 in WF 6 and 1/2 in H 2 . The activation energy was equal to 64 kJ mol -1 . An analysis of available reaction paths revealed that either the formation or the desorption of HF is the rate-limiting step. Since only one activation energy is present, the rate-limiting step is valid for the complete regime that was investigated.

The influence of adding dichlorosilane (SiH2Cl2) or germane (GeH4) to the SiH4-based reduction reaction of tungsten-hexafluoride (WF6) has been investigated in order to enhance the properties of the silane W-CVD process: e.g. selectivity and step-coverage. It is shown that the kinetics of the silane-dichlorosilane process can be characterised by a surface reaction limitation, thus improving the control of the process. For the mixed SiH4GeH4 reduction reaction of WF6 it is shown that the GeH4 process dominates the reaction kinetics. Also addition of SiH2Cl2 to the GeH4-based reduction of WF6 has been examined. It is found that in this case the formation of W is only slightly influenced. The reaction kinetics are similar to that of the unmodified deposition process.

Electromigration and stress migration in Al metallization are major reliability issues for advanced IC's. Recently it has been shown that, compared to AlSiCu alloy films, alloys containing Si, V and Pd combine excellent plasma etchability with good corrosion resistance, while a high resistance against electromigration is maintained [1]. It is commonly accepted that the resistance against stress migration, i.e. the creep strength, of Al can be improved by addition of alloying elements in combination with appropriate heat treatments (e.g. precipitation hardening). We present data on the influence of alloying elements on the behaviour of stress as a function of temperature for a number of Al-alloy films. Pure mono- and polycrystalline Al, AlSi(1.0 at%)Cu(1.0 at%) and AlV(0.1 at%)Pd(0.1 at%) films were studied. The sputter conditions, the film thickness and the annealing conditions were similar to reliability tests described in the literature. These films are subjected to thermal cycles from 50 to 425°C in a vacuum furnace, while the stress behaviour was measured by means of wafer curvature measurements.

In this work, we have incorporated heaters in a MEMS device, which allow the in situ local heating of its contacting surfaces. This design offers a promising solution for MEMS devices with contacting components by preventing capillary-induced adhesion. The force of adhesion was assessed by optically measuring in-plane snap-off displacements. We were able to decrease adhesion from 500 nN to 200 nN with just one heated surface of which the temperature was set above 300 °C. The temperature should not be set too high: we observed increased adhesion due to a direct bonding process once the temperature was increased above 750 °C. Remarkably, adhesion increased by heating from room temperature to 75 °C, which is attributed to more water being transferred to the contact area due to faster kinetics. We observed the same effect in the cases where both surfaces were heated, although at slightly different temperatures. We demonstrated that heating only one surface to between 300 °C and 750 °C is sufficient to significantly lower adhesion, due to the removal of capillary menisci. The required heater is typically most easily implemented in a stationary part of the device.

The reaction of WF 6 with a Si substrate plays an important role in the metallization of Si active areas by tungsten chemical vapor deposition (W-CUD). At typical low pressure W-CUD conditions, this reaction leads to self-limiting growth of a W film. The reaction, which takes place at the film surface, is maintained by Si diffusion through the W film, and stops when the (rapid) supply of silicon becomes rate limiting. Although the mechanism controlling the reduction of WF 6 by intrinsic Si is reasonably understood, little is known about the influence of doping atoms in the Si substrate. In this paper we describe the effect of doping atoms on the growth kinetics of tungsten films deposited by the reaction of WF 6 with highly doped Si at temperatures ranging from 275 to 360 o C. Doping atoms can have a large effect on the self-limiting time of the WF 6 /Si reaction, the self-limiting thickness of the W film, and the rate of the WF 6 /Si reaction. Similar to undoped Si substrates, depositions on As-doped n + Si are controlled by WF 6 gas-phase diffusion and the self-limiting effect occurs earlier than on undoped Si; the W growth rate on N + Si is linear in WF 6 pressure and almost independent of temperature. On B doped p + Si the WF 6 /Si reaction rate is decreased to such an extent that this surface reaction becomes rate limiting, instead of the WF 6 gas-phase diffusion; the W growth rate on p + Si is thermally activated and independent of the WF 6 pressure. The results obtained with the doped Si substrates are discussed in the light of a model for self-limiting W film growth on undoped Si substrates

When spatial objects are coated by plasma vapor deposition (PVD) or reactive PVD the objects perform a planetary motion in front of one or more targets. Sputtering from the target is achieved by an intense ion bombardment from a plasma in front of the target. This plasma is concentrated in front of the target by a magnet assembly behind the target. Usually a bias voltage is applied to the substrate in order to obtain an ion bombardment during the deposition. The function of this ion bombardment is to supply energy to the growing film in order to arrive at dense films under compressive stress. In the present article it is shown that for high quality coatings on spatial objects it is mandatory to guarantee not only a bias voltage, but also a bias current during all times of the deposition. For the coating of spatial objects this is realized by using a second plasma source to fill all of the deposition chamber with plasma.

We developed a bulge-test setup which enables measurement of the elastic and plastic properties of free-standing thin film samples between room temperature and at least 300 °C. Mechanical stress is applied to the film by a differential gas pressure across the sample and the bulge height is measured by a scanning laser beam technique. To prevent sample oxidation the pressure cell containing the sample is mounted in a vacuum chamber. The correct operation of the setup is demonstrated by measurement of the thermal expansion of free-standing Al films. Creep experiments and tensile tests demonstrate measurement of the plastic deformation of these films at temperatures up to 200 °C.

Measurements and calculations of the influence of temperature gradients on the partial pressures of the gas species in a cold‐wall chemical‐vapor‐deposition reactor are presented. The experiments were performed at low pressures (300–500 Pa total pressure) and gas mixtures consisting of hydrogen, nitrogen, and tetrafluoromethane. The partial pressures were determined by Raman spectroscopy. The Soret effect (or thermal diffusion) has a large influence on the partial pressures of heavy gases in the vicinity of the heated wafer. In some cases a decrease in partial pressure of 20% compared to the inlet partial pressures was observed. Numerical calculations were performed to predict the behavior of the gas mixture. For mixtures under investigation the gas temperatures as well as the changes in partial pressures due to the Soret effect were predicted correctly.

The step coverage of tungsten films deposited by the germane reduction of WF 6 is investigated. The step coverage modulus for the germane process is influenced only by the deposition temperature. It hardly depends on the partial pressures of the reactive species. These results are confirmed by experiments on structures with a large internal surfaces as compared to the access area. The stepcoverage of the films deposited by this process is excellent and sufficient to fill small (0.5 μm) contact holes with a high aspect ratio.

Very thin amorphous W0.7Ge0.3 films have been deposited on silicon substrates by a novel chemical vapour deposition process. For films with a thickness (d) smaller than 10 nm, the observed increase of the resistivity with decreasing film thickness is well described by percolation theory. When d >10 nm, the resistivity is independent of the film thickness ( ϱ 0 =191 μΩ cm ). The superconducting transition temperature (Tc) of thick films (d >95 nm) is 4.9 K. The observed reduction of Tc with decreasing film thickness (d 95 nm) compare very well with the result obtained by Kondo for 500 nm thick amorphous W0.7Ge0.3 films made by sputtering. For the application of amorphous W0.7Ge0.3 films in submicron superconducting devices, the excellent conformal growth of CVD films might be an important advantage.

In this paper we present in situ measurements of the wafer curvature during nucleation and growth of W films deposited by chemical vapour deposition (CVD) on Si substrates. Because of the absence of stress relaxation mechanisms in the films, these measurements directly reflect the growth stress in the W films. The growth stress development is measured during self-limiting W film growth by the Si(100) reduction of WF6 and during continuous W film growth by the H2, SiH4, and GeH4 reduction of WF6. It is shown that these measurements provide detailed information on the growth kinetics (for the Si reduction of WF6) as well as the (high) growth stresses themselves. High stress gradients are observed in the growth direction of the films. In general the growth stresses range from highly tensile during the initial stages of growth, to compressive in an intermediate region and tensile again in the thick film regime. The average film stress decreases with increasing deposition temperature. The process-dependent growth stresses reported in this paper can be used to engineer the magnitude of the average film stress in W CVD metallization of integrated circuits.

Young’s modulus of free-standing polycrystalline Al, Au, and W films with submicron thickness has been studied using a dynamic bulge-testing technique. For Au and Al films a clear frequency dependence of the modulus is observed at room temperature in the range 1×10−4–0.5 rad/s. The values of the moduli are considerably smaller than the corresponding values of bulk material. The modulus of W films measured under the same conditions does not depend on frequency and is equal to the bulk value. The origin of the behavior found in the Al and Au films is anelastic grain boundary sliding. As a consequence of the relatively small grain size of thin polycrystalline films this phenomenon is observable at room temperature in films with a relatively low melting point.

W deposition by GeH 4 reduction of WF 6 is a promising alternative for W deposition from H 2 /WF 6 and SiH 4 /WF 6 The structure and composition of W layers deposited from WF 6 and GeH 4 , are determined mainly by the deposition temperature. At temperatures between 300 and 400 o C, W layers with the A15 bcc β-W crystal structure are formed. These β-W layers contain a substantial amount of homogeneously distributed Ge, roughly between 10 and 15 atomic percent (a/o). At higher temperatures , films are formed which consist of a mixture of β-W and α-W. At temperatures greater than 500°C, the layer consists exclusively of α-W, and the Ge concentration is less than 1 a/o.

W deposition by GeH 4 reduction of WF 6 is a promising alternative for W deposition from H 2 /WF 6 and SiH 4 /WF 6 The structure and composition of W layers deposited from WF 6 and GeH 4 , are determined mainly by the deposition temperature. At temperatures between 300 and 400 o C, W layers with the A15 bcc β-W crystal structure are formed. These β-W layers contain a substantial amount of homogeneously distributed Ge, roughly between 10 and 15 atomic percent (a/o). At higher temperatures , films are formed which consist of a mixture of β-W and α-W. At temperatures greater than 500°C, the layer consists exclusively of α-W, and the Ge concentration is less than 1 a/o.

The reaction kinetics of tungsten deposition from a SiH4/WF6 gas mixture was investigated for SiH4/WF6 feed‐gas flow ratios lower than unity. By performing selective depositions, care was taken to ensure that the kinetic data was obtained in the surface‐kinetics controlled regime. It was found that the reaction kinetics of the deposition process depends on the SiH4/WF6 ratio. For low SiH4/WF6 ratios (smaller than ≊0.7) the growth rate is slightly negative in [WF6] and almost quadratic in [SiH4]. The growth rate is almost independent of temperature between 200 and 350 °C. On the basis of our measurements it is argued that in this regime the surface is covered mainly with W, WFx, and F, and that the rate‐limiting step is the dissociative adsorption of SiH4 on this surface. For higher SiH4/WF6 ratios (0.7≤SiH4/WF6≤1) the growth rate increases sharply as the SiH4/WF6 ratio increases. In this regime the growth rate cannot be expressed in the simple form: rate=A⋅[WF6]α⋅[SiH4]β since the growth rate is more sensitive to the SiH4/WF6 partial pressure ratio than to the absolute values of the SiH4 and WF6 partial pressure. For SiH4/WF6 ratios as high as 1, the growth rate is limited only by the flux of the species hitting the surface.

The growth kinetics and inhibition of growth of chemical vapor deposited thin W films on Si(100) from WF6 was studied with in situ growth stress and reflectivity measurements and ex situ weight gain measurements. A systematic series of experiments at varying WF6 flow, total pressure, and temperature show that the thickening kinetics and inhibition of the growth are controlled by two processes: WF6 diffusion through the gas phase and Si diffusion through the thickening columnar film. The steady state growth kinetics are controlled by WF6 diffusion in the gas phase whereas inhibition of the growth occurs at the transition from WF6 gas diffusion limited to Si solid state diffusion limited growth. A simple model based on WF6 gas phase diffusion and Si solid state diffusion is presented which gives a quantitative description of the experimental results.

A method for the in situ measurement of mechanical stress in thin films deposited in a vacuum system is presented. The bending of the substrate, a measure for mechanical stress in the deposited layer, is detected by reflecting two parallel laser beams off the surface of the substrate and measuring the angle between the two reflected beams. A hollow mirror in the path of the reflected beams acts as an ‘‘optical cantilever’’ and increases the sensitivity of this method. In the present setup it is possible to detect the difference between a flat substrate and a substrate with a radius of curvature of 6.5 km.

In this paper we report on in-situ stress measurements during rapid thermal annealing of titanium-silicon multilayers. After formation of the metastable TiSi 2 C49 phase relaxation of the stress in the thin film is observed. The relaxation rate of the TiSi 2 C49 phase is dependent on the overall composition of the Ti-Si multilayer. The relaxation rate for Si/Ti>2 is considerably larger than for Si/Ti

Ultra-thin amorphous WGex layers were deposited by the GeH4 reduction of WF6. In-situ reflection measurements, resistance and RBS measurements consistently showed that a closed film is formed at a thickness of about 3 nm and that subsequent growth occurs layer-by-layer with a constant growth rate. From the measured time-dependent layer thickness, reflectivity and resistivity during the growth of the 3 nm thin film it is speculated the growth starts by nucleation and growth of WGex islands. In contrast to films deposited by the H2 reduction of WF6 the Si substrate is not consumed during the GeH4 reduction of WF6. Preliminary results indicate this is caused by a Ge layer which is formed during the initial stage of the growth. In comparison with other W-LPCVD processes the GeH4 reduction of WF6 is extremely reproducible.

In this paper a thermodynamic and kinetic study of the deposition of tungsten on silicon (100) from tungsten hexafluoride (WF6) and germane (GeH4) is presented. Thermodynamic calculations, as well as experiments with a closed reactor, indicate that the reaction occurring during deposition is WF6+3GeH4→W+3GeF2+6H2. The growth rate as a function of process parameters is obtained for depositions in the temperature range from 600 to 800 K and a total pressure range from 150 to 1000 mTorr. Experiments show that the germane reduction of tungsten hexafluoride is of 0.9 order in WF6, −0.2 order in GeH4, and zero order in H2. The activation energy is 34 kJ/mol. The deposition rate does not change when SiH4 is added to the GeH4/WF6 mixture, while, on the contrary, a small amount of GeH4 reduces the growth rate from a SiH4/WF6 mixture considerably. The kinetic data indicate that the formation of GeF2 might be the rate‐limiting step.

An experimental study of etching mechanisms in Cl2 reactive ion etching is reported. The ion‐impact energy distribution and ion current density have been measured in situ at the rf electrode of a parallel‐plate reactive ion etcher. This diagnostic method has been used for the first time to unravel etch characteristics in a practical etching environment. Samples of Si, SiO2, Ti, and TiSi2 have been etched both in Cl2 and Ar discharges, and the etch rates have been related to the ion flux and impact energy distribution. The angular impact energy distribution for both ions and neutrals has been calculated numerically in order to study the contribution of fast neutrals and the angular distribution of impinging species to the etch rate. Sputter yields have been determined from the observed etch rate and the ion current density, taking into account the (angular) energy distributions of bombarding ions and neutrals. Comparison of the obtained sputter yields in Ar and Cl2 discharges with corresponding data from (...

A quantitative experimental study on the etching mechanisms of Si, SiO 2 , Ti, and TiSi 2 in Cl 2 reactive ion etching is reported. The ion impact energy distribution and the ion current density have been measured in situ at the rf-electrode of a parallel plate reactor, together with the relevant etch rates. The ion transport through the collisional rf-sheath has been simulated using a Monte Carlo method. A separate method has been used to calculate the energy distribution of bombarding neutrals. This way the contribution of ions and fast neutrals to the sputter yield in Cl 2 RIE plasmas was established. The quantitative approach has been tested against etch results in Ar plasmas in which merely physical sputtering occurs and for which well-defined experimental results are available from the literature.

An experimental study on the ion impact energy distribution and the total ion flux at the driven electrode of a parallel plate reactive ion etcher is presented. Results are shown for 13.56 MHz discharges in Ar, Ar/H2, N2, O2, Cl2, and SF6/He over a pressure range of 0.3–40 Pa. The ion impact energy distribution consists of a collision‐free part and a collision‐induced part. It is observed that in Ar, N2, O2, and Cl2 the collision‐induced part contains single and double peaks at regular energy intervals. This peaked structure is attributed to charge exchange processes in the sheath. Both the collision‐free part and the collision‐induced part of the ion impact energy distribution are well described by a model based on a constant sheath width, a sinusoidal sheath voltage, and a power law for the electric field in the sheath. The only adjustable parameter in the model is the sheath thickness. The sheath thickness has also been determined independently from the total ion current density using the Child–Langmuir law. The obtained values are in good quantitative agreement with each other and with the observed optical sheath thickness, demonstrating the overall consistency of the present approach.

The influence of adding dichlorosilane (SiH 2 Cl 2 ) to the silane (SiH 4 )-based reduction reaction of tungsten hexafluoride (WF 6 ) has been investigated to enhance the properties of this chemical vapor deposition process, especially the control of the growth rate. The growth rate of tungsten by the reduction reaction based on SiH 4 and SiH 2 Cl 2 has been measured. It is shown that the kinetics of the silane-dichiorosilane process can he characterized by a surface reaction limitation. The sole effect of the SiH 2 Cl 2 in the process is blocking of surface sites by preferential absorption, thereby reducing the growth rate of the tungsten film.

In the absence of thermal stress, tensile stress in hard metal films is caused by grain boundary shrinkage and compressive stress is caused by ion peening. It is shown that the two contributions are additive. Moreover tensile stress generated at the grain boundaries does not relax by ion bombardment. In polycrystalline hard metal films the grain structure evolves during growth, leading to wider grains higher up in the film. The tensile component of the stress in the film is generated at the grain boundaries and therefore depends on film thickness. The effect of ion bombardment is independent of grain size, therefore compressive stress does not depend on film thickness. As a result in polycrystalline films deposited under a bias voltage a stress gradient exists from tensile at the interface to compressive at the top of the film.

Thin films on substrates are usually in a stressed state. An important, but trivial, contribution to that stress stems from the difference in thermal expansion coefficient of substrate and film. Much more interesting are the intrinsic stresses, resulting from the growth and/or microstructure of the film. Intrinsic compressive stress was explained by d’Heurle in 1970. Intrinsic tensile stress for recrystallizing metal films was treated succesfully by Doljack and Hoffman in 1972. In the present letter we explain the occurrence of tensile stress in nonrecrystallizing metal films. The explanation is based on modern grain growth models and accurate stress measurements. The key ingredient to the explanation is the proof of the existence of a stress gradient in nonrecrystallizing metal films.

Hard CrNx coatings of various composition were deposited on hot work tool steel by means of reactive sputter deposition in an industrial PVD reactor. All deposited coatings (thickness 3–4 micron) were under compressive stress. The coatings were subjected to scratch testing with a diamond stylus. Two critical loads were determined. The first critical load, Lc1, indicates the onset of chipping and/or spalling. The second critical load, Lc2, indicates complete coating removal inside the track. Micrographs showing the coating failure are presented. Stoichiomeric CrN coatings outperform understoichiometric CrN0.6 coatings in scratch tests. It was found that for all specimens the coating-to-substrate adhesion is very good, independent on composition. Although the values of Lc1 and Lc2 are distinctly different, their dependence on composition is similar.

The precipitate formation occurring in Al-Cu thin foils with copper concentrations of either 1.15 at.% or 0.3 at.%, has been studied. In-situ X-ray diffraction analysis and differential scanning calorimetry are applied to determine the phases formed and the enthalpy changes in the same samples. Both X-ray diffraction and differential scanning calorimetry indicate that the precipitation behaviour of thin films (about 500 nm thickness) differs significantly from that of bulk material. In the films studied the precipitation of Al2Cu occurs at a much lower temperature than expected on the basis of the (bulk) phase diagram. Moreover, no intermediate phases are observed prior to Al2Cu precipitation. Also the amount of Cu in solid solution (0.20 at%Cu) observed by electron-probe micro-analysis after slowly cooling from 500°C to room temperature, exceeds the solubility of bulk Al-Cu.

The influence of Cu on the response of Al-Cu thin films to thermally induced stress is studied. The copper concentration is varied between 0 and 1.15 at. %. It is proposed that copper atoms which have not formed precipitates, largely affect the mechanical behaviour. This idea is supported by the following observations. An isothermal hold results in temporary strengthening of the films. The extent of this strengthening increases with copper concentration, increases with decreasing isothermal hold temperature and saturates with increasing isothermal hold period. Based on these observations the large tensile stress increase below 200 °C is ascribed to the formation of Cottrell atmospheres.