Abstract: The authors report a new chemical approach for the selective atomic layer deposition of ultrathin layers of zirconium oxide (ZrO2) on copper patterned silicon surfaces. Instead of using common atomic layer deposition (ALD) oxygen sources such as water, oxygen, or ozone, the authors use ethanol, which serves as oxygen source for the ALD on the silicon side and as effective reducing agent on the copper side, thereby selectively depositing ZrO2 film on the silicon surface of the substrate without any deposition on copper up to at least 70 ALD cycles. The resulting ZrO2 nanofilm is found to be an effective copper diffusion barrier at temperatures at least up to 700 °C.

Abstract: Diffusion is one of the fundamental processes that govern the structure, processing, and properties of materials and it plays a crucial role in determining device lifetimes. However, direct observations of diffusion processes have been elusive and limited only to the surfaces of materials. Here we use an aberration-corrected electron microscope to locally excite and directly image the diffusion of single Ce and Mn dopants inside bulk wurtzite-type AlN single crystals, identifying correlated vacancy-dopant and interstitial-dopant kick-out mechanisms. Using a 200 kV electron beam to supply energy, we observe a higher frequency of dopant jumps for the larger and heavier Ce atoms than the smaller Mn atoms. These observations confirm density-functional-theory-based predictions of a decrease in diffusion barrier for large substitutional atoms. The results show that combining depth sensitive microscopy with theoretical calculations represents a new methodology to investigate diffusion mechanisms, not restricted to surface phenomena, but within bulk materials.

Abstract: Protective caps residing at an interface between copper lines and dielectric diffusion barrier layers are used to improve various performance characteristics of interconnects. The caps, such as cobalt-containing caps or manganese-containing caps, are selectively deposited onto exposed copper lines in a presence of exposed dielectric using CVD or ALD methods. The deposition of the capping material is affected by the presence of carbon-containing contaminants on the surface of copper, which may lead to poor or uneven growth of the capping layer. A method of removing carbon-containing contaminants from the copper surface prior to deposition of caps involves contacting the substrate containing the exposed copper surface with a silylating agent at a first temperature to form a layer of reacted silylating agent on the copper surface, followed by heating the substrate at a higher temperature to release the reacted silylating agent from the copper surface.

Abstract: The evolution of copper-based interconnects requires the realization of an ultrathin diffusion barrier layer between the Cu interconnect and insulating layers. The present work reports the use of atomically thin layer graphene as a diffusion barrier to Cu metallization. The diffusion barrier performance is investigated by varying the grain size and thickness of the graphene layer; single-layer graphene of average grain size 2 ± 1 μm (denoted small-grain SLG), single-layer graphene of average grain size 10 ± 2 μm (denoted large-grain SLG), and multi-layer graphene (MLG) of thickness 5-10 nm. The thermal stability of these barriers is investigated after annealing Cu/small-grain SLG/Si, Cu/large-grain SLG/Si, and Cu/MLG/Si stacks at different temperatures ranging from 500 to 900 °C. X-ray diffraction, transmission electron microscopy, and time-of-flight secondary ion mass spectroscopy analyses confirm that the small-grain SLG barrier is stable after annealing up to 700 °C and that the large-grain SLG and MLG barriers are stable after annealing at 900 °C for 30 min under a mixed Ar and H2 gas atmosphere. The time-dependent dielectric breakdown (TDDB) test is used to evaluate graphene as a Cu diffusion barrier under real device operating conditions, revealing that both large-grain SLG and MLG have excellent barrier performance, while small-grain SLG fails quickly. Notably, the large-grain SLG acts as a better diffusion barrier than the thicker MLG in the TDDB test, indicating that the grain boundary density of a graphene diffusion barrier is more important than its thickness. The near-zero-thickness SLG serves as a promising Cu diffusion barrier for advanced metallization.

Abstract: An interconnect structure located on a semiconductor substrate within a dielectric material positioned atop the semiconductor substrate is provided having an opening within the dielectric material, the opening includes an electrically conductive material extending from the bottom to the top, and contacting the sidewall; a first layer located on the sidewall of the opening, the first layer is made from a material including titanium oxide or titanium silicon oxide; a second layer located between the first layer and the electrically conductive material, the second layer is made from a material selected from the group TiXO b , TiXSi a O b , XO b , and XSi a O b , X is Mn, Al, Sn, In, or Zr; and a third layer located along a top surface of the electrically conductive material, the third layer is made from a material selected from the group TiXO b , TiXSi a O b , XO b , and XSi a O b , X is Mn, Al, Sn, In, or Zr.

Abstract: We demonstrate the thinnest ever reported Cu diffusion barrier, a 1-nm-thick graphene tri-layer. X-ray diffraction patterns and Raman spectra show that the graphene is thermally stable at up to 750 °C against Cu diffusion. Transmission electron microscopy images show that there was no inter-diffusion in the Cu/graphene/Si structure. Raman analyses indicate that the graphene may have degraded into a nanocrystalline structure at 750 °C. At 800 °C, the perfect carbon structure was damaged, and thus the barrier failed. The results of this study suggest that graphene could be the ultimate Cu interconnect diffusion barrier.

Abstract: Nanocomposite films made up of either Pt–Rh/ZrO2 or Pt–Rh/HfO2 materials were co-deposited using multiple e-beam evaporation sources onto langasite (La3Ga5SiO14) substrates, both as blanket films and as patterned interdigital transducer electrodes for surface acoustic wave sensor devices. The films and devices were tested after different thermal treatments in a tube furnace up to 1,200 °C. X-ray diffraction and electron microscopy results indicate that Pt–Rh/HfO2 films are stabilized by the formation of monoclinic HfO2 precipitates after high temperature exposure, which act as pinning sites to retard grain growth and prevent agglomeration of the conductive cubic Pt–Rh phase. The Pt–Rh/ZrO2 films were found to be slightly less stable, and contain both tetragonal and monoclinic ZrO2 precipitates that also helps prevent Pt–Rh agglomeration. Film conductivities were measured versus temperature for Pt–Rh/HfO2 films on a variety of substrates, and it was concluded that La and/or Ga diffusion from the langasite substrate into the nanocomposite films is detrimental to film stability. An Al2O3 diffusion barrier grown on langasite using atomic layer deposition was found to be more effective than a SiAlON barrier layer in minimizing interdiffusion between the nanocomposite film and the langasite crystal at temperatures above 1,000 °C.

Abstract: We report penetration and lateral diffusion behavior of environmental molecules on synthesized polycrystalline graphene. Penetration occurs through graphene grain boundaries resulting in local oxidation. However, when the penetrated molecules diffuse laterally, the oxidation region will expand. Therefore, we measured the lateral diffusion rate along the graphene–copper interface for the first time by the environment-assisted crack growth test. It is clearly shown that the lateral diffusion is suppressed due to the high van der Waals interaction. Finally, we employed bilayer graphene for a perfect diffusion barrier facilitated by decreased defect density and increased lateral diffusion path.

Abstract: Provided are resistive random access memory (ReRAM) cells and methods of fabricating thereof. A ReRAM cell includes an embedded resistor and a resistive switching layer connected in series with this resistor. The resistor is configured to prevent over-programming of the cell by limiting electrical currents through the resistive switching layer. Unlike the resistive switching layer, which changes its resistance in order to store data, the embedded resistor maintains a substantially constant resistance during operation of the cell. The embedded resistor is formed from tantalum nitride and silicon nitride. The atomic ratio of tantalum and silicon may be specifically selected to yield resistors with desired densities and resistivities as well as ability to remain amorphous when subjected to various annealing conditions. The embedded resistor may also function as a diffusion barrier layer and prevent migration of components between one of the electrodes and the resistive switching layer.

Abstract: This paper presents the use of graphene as a diffusion barrier to a eutectic Ga-In-Sn alloy, i.e., galinstan, for electrical contacts in electronics. Galinstan is known to be incompatible with many ...

Abstract: By means of density functional theory computations, we systematically investigated the behavior of lithium (Li) adsorption and diffusion on MoO3 with different dimensions: including three-dimensional (3D) bulk, two-dimensional (2D) double-layer, 2D monolayer and one-dimensional (1D) nanoribbons. The Li binding energies and diffusion barriers are comparable in MoO3 bulk and double-layer. Reducing the dimension to the MoO3 monolayer simultaneously lowers the Li diffusion barrier and the interaction between Li atoms and the MoO3 monolayer. Cutting the MoO3 monolayer into 1D nanoribbons can further facilitate the diffusion of Li atoms, and enhance the Li binding energies. Especially, Li diffusion on nanoribbons is rather facile along both the axial and the transverse directions. These computational results demonstrate that due to the dimensional reduction, MoO3 monolayer nanosheets and nanoribbons have exceptional properties (good electronic conductivity, fast Li diffusion, high operating voltage and high energy density), and thus are promising as high-rate Li ion battery electrodes.

Abstract: In order to study the influence of the metallic substrate on the catalytic activity of structured micromonolithic catalysts, a CuO x /CeO 2 catalyst was deposited on different oxidized or enameled metallic micromonoliths and tested in the PROX reaction under ideal and realistic conditions. The obtained results show as both activity and selectivity depend on the nature of the alloy and the nature of the interphase between the metal substrate and the catalyst layer. In oxidized micromonoliths, diffusion of Cr and Fe has been observed. For enameled micromonoliths, together with that diffusion, the interaction of the glass–ceramic interphase with the reactive gas streams resulted in the partial hydrolysis of this layer leading to diffusion toward the catalyst surface of the hydrolysis products, namely Na, Ca and Si cations. In some cases, the alteration of the surface composition favors the spreading of the copper active phase. As a result, it must be concluded that the metallic substrates are not spectators, at least in the PROX reaction, playing a fundamental role in the performances of the catalytic devices.

Abstract: A package component includes a surface dielectric layer having a first planar surface, and a metal pad in the surface dielectric layer. The metal pad includes a diffusion barrier layer that includes sidewall portions, and a metallic material encircled by the sidewall portions of the diffusion barrier layer. The metallic material has a second planar surface level with the first planar surface. An air gap extends from the second planar surface of the metallic material into the metallic material. An edge of the air gap is aligned to an edge of the metallic material.

Abstract: This study was performed in order to investigate a possibility to obtain Co-W microbumps via electrochemical routes, because this alloy recently has gained attraction as a novel barrier against copper diffusion. In order to be applied in flip-chip technology, barrier layers should be void-free and uniformly deposited on the entire area of a die to ensure high reliability and high performance of wafer bump-solder interface. To meet these requirements, a set of potentiostatic and galvanostatic electrodeposition was carried out from a citrate electrolyte, at pH 5 and at room temperature. The tests done confirm that void-free Co-W bumps with a uniform tungsten content along the bump can be obtained by potentiostatic and galvanostatic electrodeposition. Successful electrodeposition of Cu/Co-W/Sn layers with good adhesion between them and uniformity on the entire array of bumps also was obtained. The XPS data confirm that electrodeposited Co-W layers can act as a good barrier between Sn and Cu.

Abstract: Microstructure and composition analysis was performed on the nitrided Ti–6Al–4V alloy by using coupled techniques of X-ray diffraction analysis and glow discharge spectroscopy. Titanium nitride (TiN x ) appeared at the outmost surface, but the low solubility of Al in the TiN x phase made Al atom to move inward. Such Al-rich phase as Ti 3 Al compound was formed by Al accumulation at the boundary between TiN x and α(N)-Ti phase (N diffusion zone). Interestingly, the Ti 3 Al had very low concentration of N, and it may inhibit N atom from diffusing inward as a diffusion barrier. The flat region of V concentration profile with about 2 at% also appeared in the region of α(N)-Ti phase because there was a solubility limit of less than 2 at% V in the α-Ti phase (hcp).

Abstract: Provided are resistive random access memory (ReRAM) cells having diffusion barrier layers formed from various materials, such as beryllium oxide or titanium silicon nitrides. Resistive switching layers used in ReRAM cells often need to have at least one inert interface such that substantially no materials pass through this interface. The other (reactive) interface may be used to introduce and remove defects from the resistive switching layers causing the switching. While some electrode materials, such as platinum and doped polysilicon, may form inert interfaces, these materials are often difficult to integrate. To expand electrode material options, a diffusion barrier layer is disposed between an electrode and a resistive switching layer and forms the inert interface with the resistive switching layer. In some embodiments, tantalum nitride and titanium nitride may be used for electrodes separated by such diffusion barrier layers.

Abstract: A remarkable enhancement of atomic diffusion is highlighted by scanning tunneling microscopy performed on ultrathin metastable body-centered tetragonal Co films grown on Fe(001). The films follow a nearly perfect layer-by-layer growth mode with a saturation island density strongly dependent on the layer on which the nucleation occurs, indicating a lowering of the diffusion barrier. Density functional theory calculations reveal that this phenomenon is driven by the increasing capability of the film to accommodate large deformations as the thickness approaches the limit at which a structural transition occurs. These results disclose the possibility of tuning surface diffusion dynamics and controlling cluster nucleation and self-organization.

Abstract: U-Mo alloys are being developed as low enrichment monolithic fuel under the Reduced Enrichment for Research and Test Reactor (RERTR) program. Diffusional interactions between the U-Mo fuel alloy and Al-alloy cladding within the monolithic fuel plate construct necessitate incorporation of a barrier layer. Fundamentally, a diffusion barrier candidate must have good thermal conductivity, high melting point, minimal metallurgical interaction, and good irradiation performance. Refractory metals, Zr, Mo, and Nb are considered based on their physical properties, and the diffusion behavior must be carefully examined first with U-Mo fuel alloy. Solid-to-solid U-10 wt.%Mo versus Mo, Zr, or Nb diffusion couples were assembled and annealed at 600, 700, 800, 900 and 1000 °C for various times. The interdiffusion microstructures and chemical composition were examined via scanning electron microscopy and electron probe microanalysis, respectively. For all three systems, the growth rate of interdiffusion zone were calculated at 1000, 900 and 800 °C under the assumption of parabolic growth, and calculated for lower temperature of 700, 600 and 500 °C according to Arrhenius relationship. The growth rate was determined to be about 103 times slower for Zr, 105 times slower for Mo and 106 times slower for Nb, than the growth rates reported for the interaction between the U-Mo fuel alloy and pure Al or Al-Si cladding alloys. Zr, however was selected as the barrier metal due to a concern for thermo-mechanical behavior of UMo/Nb interface observed from diffusion couples, and for ductile-to-brittle transition of Mo near room temperature.