Surface Of Cu Film Research Articles

High-Efficiency Copper Removal by Nitrogen Plasma-Assisted Picosecond Laser Processing.

Copper (Cu) removal efficiency is a key parameter in the processing of Cu-based electronic devices. Herein, a nitrogen plasma-assisted picosecond (ps) laser process for Cu removal is presented. Based on the cleaning and activation effect of nitrogen plasma on the surface of Cu film in ps-laser ablation, the removal efficiency can be significantly improved. Theoretically, the interaction mechanism between Cu and the ps-laser under the action of the plasma flow field is investigated by the dual temperature model (TTM) and finite element analysis (FEA). Meanwhile, the experimental results show that the angle of the plasma flow significantly affects the laser ablation of Cu. Small-angle plasma helps to improve the ps-laser processing precision of Cu, while large-angle plasma can effectively improve the ps-laser processing efficiency of Cu. Under the laser fluence of 2.69 J/cm2, the removal depth of the Cu film by a 30° plasma-assisted ps-laser is 148% higher than that by the non-plasma-assisted ps-laser, which indicates the application potential of nitrogen plasma in improving the laser ablation process.

Study on the film forming mechanism, corrosion inhibition effect and synergistic action of two different inhibitors on copper surface chemical mechanical polishing for GLSI

As the most basic interconnection metallic material, copper (Cu) has been widely used in giant-large scale integrated circuits (GLSI). Corrosion inhibitors always play a crucial role in the Cu chemical mechanical polishing (CMP) process. The passivation of hydrogen peroxide (H2O2) alone is too weak to adequately protect against chemical corrosion on the surface of Cu film, thus, it is necessary to add corrosion inhibitors to the slurry. As a representative corrosion inhibitor, 1, 2, 4-triazole was compared with a novel corrosion inhibitor known as TT-LYK (ethanol, 2, 2′-[[(methyl-1H-benzotriazol-1-yl) methyl] imino] bis-), and their mixed inhibitor was studied in this paper. Different analytical methods such as the potentiodynamic polarization test, physical appearance phenomenon analysis, contact angle measurement and dynamic polishing experiment were used to compare their inhibition properties, and the fitting of different adsorption isotherms and static erosion tests were used to clarify the adsorption types and passivation processes respectively. The results indicated that both TT-LYK and 1, 2, 4-triazole achieved the goal of inhibiting Cu surface corrosion, but the inhibition effect of TT-LYK was slightly weaker at the same mass fraction. The corrosion protection of the two inhibitors is mainly attributable to the presence of organic inhibitor molecules on the surface of Cu by chemisorption and physisorption simultaneously. The passivation processes of both can be divided into two steps. First, the direct growth of the Cu-inhibitor passivation film occurs, followed by the redeposition of the Cu-inhibitor complex. From the results of electrochemical experiments, it can be concluded that the structure of the thin passivation film growing on the Cu surface varies with the inhibitor, and the composition of the passivation film was investigated by XPS and Raman spectrum analysis. At the same time, the surface roughness after polishing was found to decrease with the increase in inhibition effect of the inhibitors. According to the inhibition characteristics of these two inhibitors, a mixed corrosion inhibitor (TT-LYK and 1, 2, 4-triazole) was proposed in this study. Under the synergistic inhibition action, better inhibition effect and surface quality for Cu CMP were obtained. This outcome may be ascribed to the variation in the structure and density of the passivation film.

Study of Cu Film Surface Treatment Using Formic Acid Vapor/Solution for Low Temperature Bonding

In this paper, the surface of Cu film was treated with formic acid vapor and solution for Cu low temperature bonding. After formic acid vapor/solution treatment, Cu film surface oxide was reduced, and surface became rougher. For formic acid vapor treatment, with the increase of treatment time, Cu film surface was reduced gradually at 200°C, surface roughness and particle size became bigger. For formic acid solution treatment, the reduction of the surface of the copper film was the most abundant when the solution concentration is 50%. After the treatment with formic acid solution, the surface roughness and particle size of copper film increase significantly. As the formic acid solution on the surface of copper film has a certain role in corrosion, using formic acid solution treatment, better surface reduction and rougher surface for Cu film were obtained. By contrast, using formic acid vapor treatment, surface reduction is not so good, but surface is smoother than that treated by formic acid solution. Cu/Cu direct bonding was realized at 200°C after formic acid vapor/solution treatment. Using formic acid vapor treatment, Cu/Cu bond strength is about 18.2 MPa, which is higher than that using formic acid solution treatment.

The Effect of Thickness and Annealing Time on Surface Segregation in Cu/Ni, Ni/Cu Film/Substrate and Cu/Ni/Si Film/Film/Substrate

The Cu/Ni, Ni/Cu and Cu/Ni/Si film/substrate and film/film/substrate systems were prepared by magnetron sputtering method to investigate the surface segregation. The chemical composition of film was analyzed by Auger Electron Spectroscopy (AES). The microtopographies of the Cu/Ni surface and the cross section of the film were observed by Transmission Electron Microscope (TEM), where the blocky distribution of Ni-rich area on surface of Cu film and columnar grains was observed in the specimen. It is found that the thickness of sputtered film has stronger effect on the composition of segregation layer near the surface than that of the annealing time. The surface segregation could be ascribed to the fast vertical diffusion of the substrate atoms through the columnar grain boundaries and the subsequent lateral surface diffusion.

Significant surface flattening effect by Au addition for Cu growth on Cu3Au(001)

To prepare a flat surface of Cu film on Cu3Au(001), we utilized the effect of Au addition to Cu, and investigated the dependence of the growth mode on the amount of Au added. We grew Cu–x%Au (x=0–20) films on Cu3Au(001) underlayers by co-deposition and observed the surface morphology by scanning tunneling microscopy. For Cu film without Au addition, three-dimensional islands were observed on the surface while flat two-dimensional surfaces were observed for Cu film with only 2.5%Au addition. This difference in the growth mode was found to be due to surface segregation of Au revealed by in situ Auger electron spectroscopy. It was considered that the Au atoms acted like a surfactant and avoided three-dimensional island growth by suppression of the accumulation of epitaxial strain in Cu–x%Au films. However, the terrace size decreased with x. This decrease was thought to be due to an increase in an effective Ehrlich–Schwoebel barrier. The surface flattening effect by addition of another element would be obtained in other systems as well by employing a suitable element.

Cleaning Effects by NH₄OH Solution on Surface of Cu Film for Semiconductor Devices

We investigated cleaning effects using <TEX>$NH_4OH$</TEX> solution on the surface of Cu film. A 20 nm Cu film was deposited on Ti / p-Si (100) by sputter deposition and was exposed to air for growth of the native Cu oxide. In order to remove the Cu native oxide, an <TEX>$NH_4OH$</TEX> cleaning process with and without TS-40A pre-treatment was carried out. After the <TEX>$NH_4OH$</TEX> cleaning without TS-40A pretreatment, the sheet resistance Rs of the Cu film and the surface morphology changed slightly(<TEX>${\Delta}Rs:{\sim}10m{\Omega}/sq.$</TEX>). On the other hand, after <TEX>$NH_4OH$</TEX> cleaning with TS-40A pretreatment, the Rs of the Cu film changed abruptly (<TEX>${\Delta}Rs:till{\sim}700m{\Omega}/sq.$</TEX>); in addition, cracks showed on the surface of the Cu film. According to XPS results, Si ingredient was detected on the surface of all Cu films pretreated with TS-40A. This Si ingredient(a kind of silicate) may result from the TS-40A solution, because sodium metasilicate is included in TS-40A as an alkaline degreasing agent. Finally, we found that the <TEX>$NH_4OH$</TEX> cleaning process without pretreatment using an alkaline cleanser containing a silicate ingredient is more useful at removing Cu oxides on Cu film. In addition, we found that in the <TEX>$NH_4OH$</TEX> cleaning process, an alkaline cleanser like Metex TS-40A, containing sodium metasilicate, can cause cracks on the surface of Cu film.

Improvement of Interfacial Adhesion Strength and Thermal Stability of Cu/p-SiC:H/SiOC:H Film Stack by Plasma Treatment on the Surface of Cu Film

A highly reliable interface of self-aligned barrier CuSiN thin layer between the Cu film and the nano-porous SiC:H (p-SiC:H) capping barrier (k=3.3) has been developed in the present work. With the introduction of self-aligned barrier (SAB) CuSiN between a Cu film and a p-SiC:H capping barrier, the interfacial thermal stability and the adhesion of the Cu/p-SiC:H film are considerably enhanced. A significant improvement of adhesion strength and thermal stability of Cu/p-SiC:H/SiOC:H film stack has been achieved by optimizing the pre-clean step before cap-layer deposition and by forming the CuSiN-like phase. This cap layer on the surface of the Cu can provide a more cohesive interface and effectively suppress Cu atom migration as well.

Effects of Ti Addition on Via Reliability in Cu Dual Damascene Interconnects

We investigated the effects of a Ti addition on the reliability and the electrical performance of Cu interconnects, comparing three different ways of Ti addition such as A) Ti layer insertion under Ta-TaN stacked barrier metal, B) Ti layer insertion between a Ta-TaN barrier and Cu, and C) the Ti doping from the surface of the electrochemical-plated (ECP) Cu film. The structure-A drastically suppresses the stress-induced voiding (SIV) under the via connected to a wide lower line due to adhesion improvement by Ti at the via-bottom, while the electromigration (EM) is not improved. In the structure-B, by contrast, the EM is improved but the SIV resistance is degraded. The Ti doping from the bottom surface of Cu film restricts the grain growth and increases the tensile stress, enhancing the SIV. The structure-C improves not only the SIV but also the EM resistance. The oxygen gettering effect of Ti during the ECP-Cu annealing is a reason for the reliability improvements of the SIV and the EM. The improvement of adhesiveness at the interface between the via and the lower Cu line, and the oxygen gettering from Cu by Ti play an important role in suppressing the SIV and the EM.

Oxidized surfaces of Cu films on Ru(0001)

The crystallographic aspects of O-induced structures on strained Cu films grown epitaxially on Ru(0001) are studied by scanning tunneling microscopy and high-resolution low-energy electron diffraction. O adsorption at 520 K removes the thickness-dependent relaxation structures of the O-free Cu film. Instead, structures appear corresponding to a film thickness reduced by one monolayer, accompanied with the formation of a O/Cu surface layer. Wave-like O–Cu–O chains along the [ 1 ̄ 12] direction are formed on the pseudomorphic Cu monolayer. The O/Cu structures on uniaxially reconstructed Cu bilayers depend on both the local stacking sequence and the extension of the unit cell. On moiré-like Cu films (thickness >3 ML), an O/Cu structure of well-established long-range order only occurs if O is preadsorbed on Ru(0001). The results obtained allow one to conclude that Cu–O segments (links) of the O–Cu–O chains should be regarded as surfactants for the O-mediated growth of Cu/Ru(0001). All O/Cu structures can be interpreted on the basis of O–Cu–O strings, in close relation to Cu 2O(111).

Energetic beam-induced surface segregation during deposition

The energetic beam-induced surface segregation of metal (Cu thin film)/metal (Ag substrate) structure irradiated by energetic atom and ion beams has been observed. It has been found that Ag was segregated onto the surface of Cu film (about 15 nm in thickness deposited under UHV) during the ion beam deposition and irradiation of the deposited film afterwards. The energy of the Ar + ion beam used both for sputter-deposition and irradiation ranged over several keV. All the measurements were made with a PHI-590 scanning Auger microprobe at room temperature at a base pressure of about 6 × 10 −8 Pa. During sputter-deposition the Ag enrichment is greater the higher the energy of the sputtering ion beam used. This kinetic energy influence of deposited atom on the degree of segragation is interesting as the average energy of atoms sputtered with an ion beam energy of 2–4 keV is several eV and their energy difference should be in the range 10 −1-1 eV. The deposited film shows greatest enrichment of Ag when it is irradiated by an Ar + ion beam with keV energy.

Diffuse optical scattering from Cu surfaces with variable roughness

Polarization data from He–Cd laser light scattered from the surfaces of Cu films deposited in UHV and measured in situ, are compared with calculations to investigate if we can conclude directly from the spectra whether first order theory is applicable. The ratio of the pp and ss diffusely scattered radiation differs by as much as 20% from the values predicted for the smoothest surfaces of unannealed films for which the scattering probability per unit solid angle is ∠10−5 for k≳10−3 Å−1. When the size of the majority of the surface undulations is smaller than the wavelength of the incident light, the measured polarization ratio exceeds that calculated from first order theory. The roughness of Cu films and the diffuse scattered intensity increase by annealing to temperatures of 700°C. The discrepancy between the data and calculation increases with the roughness. For a macroscopically rough surface the experimental polarization ratio is smaller than the theoretical value. The cross polarization intensities are nonzero for all surfaces after annealing but this has not been found to be a useful indication of the limit of validity of first order scattering theory. Differences between the measured polarization ratios and those predicted by first‐order scattering theory are explained in terms of geometrical and multiple scattering effects.