Haruo Akahoshi

Electroless deposited cobalt-tungsten-boron capping barrier metal on damascene copper interconnection

The metal capping barrier deposited by the electroless cobalt tungsten boron (CoWB) alloy plating method for ultralarge scale integration applications was investigated. The CoWB film was formed directly on copper without a palladium catalyst, using dimethyl amin borane (DMAB) as a reducing agent, and it was deposited selectively on 0.25 μm wide copper interconnects separated with 0.25 μm spacing SiO 2 . The CoWB thin films were effective barriers against copper diffusion even at CoWB thicknesses as low as 50 nm. Compared with the CoWB film, cobalt tungsten phosphorus films deposited directly on copper using DMAB as a deposition initiator was not effective as a copper diffusion barrier. The plating films contained mainly cobalt with a significant amount of tungsten (up to 20 atom %) and a small amount of boron. Additionally, we propose a newly developed alkaline metal free electroless CoWB plating solution using tetramethyl ammonium hydroxide as a pH adjuster.

Electroless deposited CoWB for copper diffusion barrier metal

A metal capping barrier deposited by an electroless CoWB (cobalt tungsten boron) alloy plating method on damascene copper interconnects has been investigated. The metal capping barrier structure is one solution to prevent the decrease of coupling capacity associated with SiN or SiC capping barriers. In this paper, we propose a direct electroless deposition of Co alloy on copper surfaces, using dimethyl amine borane (DMAB) as a reducing agent, without a palladium catalyst.

Binary Mixed Solvent Electrolytes Containing Trifluoropropylene Carbonate for Lithium Secondary Batteries

To improve the safety of the electrolyte used for lithium secondary batteries, binary mixed solvent electrolytes containing trifluoropropylene carbonate (TFPC) as cosolvent have been studied. Chloroethylene carbonate (CIEC), ethylene carbonate, and propylene carbonate were chosen as the other component of the binary mixed solvent for the electrolytes. The solution properties of these electrolytes were characterized using conductivity and nuclear magnetic resonance (NMR) spectroscopy. The chemical shift of CIEC and TFPC did not vary with the mixing ratio due to their similar enthalpies of solvation as derived by molecular orbital simulation. The ClEC/TFPC electrolyte showed higher discharge capacities with lower irreversible capacity loss in both a graphite/Li cell and Li 1+x Mn 2 O 4 /Li cell than other electrolyte systems. Electrochemical impedance spectroscopy measurements were made for cells composed of each electrolyte. The surface of the graphite anode was analyzed using X-ray photoelectron spectroscopy, infrared spectroscopy, and solid 7 Li-NMR spectroscopy.

Glass-modified stress waves for adhesion measurement of ultra thin films for device applications

Abstract Laser-generated stress wave profiles with rarefaction shocks (almost zero post-peak decay times) have been uncovered in different types of glasses and presented in this communication. The rise time of the pulses was found to increase with their amplitude, with values reaching as high as 50 ns . This is in contrast to measurements in other brittle crystalline solids where pulses with rise times of 1 – 2 ns and post-peak decay times of 16 – 20 ns were recorded. The formation of rarefaction shock is attributed to the increased compressibility of glasses with increasing pressures. This was demonstrated using a one-dimensional nonlinear elastic wave propagation model in which the wave speed was taken as a function of particle velocity. The technological importance of these pulses in measuring the tensile strength of very thin film interfaces is demonstrated by using a previously developed laser spallation experiment in which a laser-generated compressive stress pulse in the substrate reflects into a tensile wave from the free surface of the film and pries off its interface at a threshold amplitude. Because of the rarefaction shock, glass-modified waves allow generation of substantially higher interfacial tensile stress amplitudes compared with those with finite post-peak decay profiles. Thus, for the first time, tensile strengths of very strong and ultra thin film interfaces can be measured. Results presented here indicate that interfaces of 185-nm-thick films, and with strengths as high as 2.7 GPa , can be measured. Thus, an important advance has been made that should allow material optimization of ultra thin layer systems that may form the basis of future MEMS-based microelectronic, mechanical and clinical devices.

Impedance of a reaction involving two adsorbed intermediates: aluminum dissolution in non-aqueous lithium imide solutions

The model presented considers the dissolution of a trivalent metal in three consecutive steps involving two adsorbed intermediates. If mass transport effects are negligible, it is possible to construct equivalent circuits in which adsorption-related elements are doubled compared to the case of a single adsorbate. In the case where mass transport affects the dissolution, the Faradaic admittance can be evaluated as a fraction of two power series and no simple equivalent circuit can be constructed from conventional circuit elements. Depending on the mechanism assumed, the low-frequency behavior can be either similar to a Warburg impedance or different fundamentally. The impedance of aluminum dissolution is discussed in the case of insignificant mass transport. The Langmuir isotherm is supposed to hold for intermediate adsorption, and only anodic partial reactions are accounted for. It has been concluded that the second step is rate-determined and that solvent takes part in the desorption of the product only. An empirical correlation was found between the dipole moment of the solvent used and the ratio of the rate constants of non-rate determining steps.

Effect of the Purity of Plating Materials on the Reduction of Resistivity of Cu wires for Future LSIs

Resistivity difference between Cu wires made with plating using high purity (new plating process) and conventional purity (conventional process) materials has been evaluated in order to develop the process for the realization of high performance LSIs. This resistivity difference is relatively small, i.e., 8% when line width is wide (200 nm). However, it increases with the decrease in line width, and it reaches about 20%, i.e., 2.8 μΩ cm for the former and 3.5 μΩ cm for the latter at 50 nm line width. A 50 nm wide Cu wire formed with the new plating process had more uniform and larger grain sizes and lower impurity concentrations than the wire formed with the conventional process.

Observation of microstructures in the longitudinal direction of very narrow Cu interconnects

We have succeeded in observing the longitudinal microstructure of very narrow Cu interconnects for the first time. We found that the average grain sizes along the longitudinal direction of Cu interconnect trenches increased with increasing line width, and they were 278 nm for 80 nm, 303 nm for 100 nm, and 346 nm for 180 nm wide interconnects. Ratios of the average grain size to line width were 3.5 for 80 nm, 3.03 for 100 nm, and 1.9 for 180 nm line widths.

Effects of Minor Elements in Cu Leadframe on Whisker Initiation From Electrodeposited Sn/Cu Coating

Drastically different tendencies of whisker initiation have been observed from the same Sn/Cu coating electrodeposited on two different Cu leadframe materials, CUFE and CUCR. After long-term storage at room temperature, no whisker initiation was observed on the coating on CUCR, whereas long whiskers with a maximum length of more than 200 mum were formed on the coating on CUFE. FE-STEM and FE-TEM microstructural observations of vertical sections of each Cu leadframe with the coating showed that the cross-sectional morphologies of Cu-Sn intermetallic compounds (IMCs) formed at the interface between the coating and the leadframe were as follows: a wedge-shaped structure for the Sn/Cu- CUFE sample and a comb-tooth structure for the Sn/Cu-CUCR sample. Energy dispersive X-ray (EDX) analysis results confirmed that the formation morphologies of the Cu-Sn IMC were affected by minor elements in the Cu leadframes. Fe particles with a large diameter of mainly 50-200 nm precipitated nonuniformly in the CUFE leadframe, which had Fe, Zn, and P as minor elements. Conversely, very small Cr-rich particles, 10-20 nm in diameter, were precipitated in the CUCR leadframe, which had Cr, Sn, and Zn as minor elements. The Fe particles suppressed the growth of the Cu-Sn IMC by acting as obstacles to IMC formation, whereas grain boundaries in the Sn/Cu coating acted as preferential IMC formation sites. The combination of these two effects is thought to produce the wedge-shaped cross-sectional structure of the IMC. On the other hand, the Cr-rich particles did not have such a prominent suppression effect on Cu-Sn IMC growth: only the grain boundaries in the Sn/Cu coating affected the IMC formation site. As a result, IMC with a comb-tooth structure was formed in the Sn/Cu- CUCR sample. The difference in whisker initiation tendency was inferred to be due to the difference in compressive stress on the basis of the Cu-Sn IMC formation morphology in the Sn/Cu coating. Therefore, the stresses were finally measured by X-ray diffraction to determine the correlation between whisker initiation tendency and the morphology of Cu-Sn IMC formation. Based on the results, a model for controlling whisker initiation in Sn/Cu-coated Cu leadframe systems is presented.

Correlation Between Whisker Initiation and Compressive Stress in Electrodeposited Tin–Copper Coating on Copper Leadframes

To evaluate the contribution of coating stress to whisker initiation from IC package leads, the stress distribution in the coating was investigated by finite-element analysis (FEA). Two different leadframe samples, which were composed of the same tin-copper coating on two different copper-leadframe materials, namely, copper-iron (hereafter, CUFE; corresponding to CDA number C19400) and copper-chromium (CUCR; CDA number C18045), were used to examine the whisker-initiation behavior on the coating surfaces. The two samples showed significantly different tendencies of whisker initiation from the coating. That is, after long-term storage at room temperature, no whisker initiation was observed on the coating on the CUCR sample, whereas long whiskers (with a maximum length of more than 200 μm) were formed from the coating on the CUFE sample. The FEA calculation on the leadframe samples revealed that the coatings had a two-directional stress gradient, namely, one gradient toward the surface and another toward the base leadframe material. It also indicated a difference between the stress distributions in the two samples. The gradient of normal stress on the coatings grain boundaries (GBs), toward the surface of the CUFE sample, was found to be larger than that in the CUCR sample. This result implies that the tin-atom flux along a GB in the coating on the CUFE sample was larger than that on the CUCR sample because the atom flux along the GB was proportional to the stress gradient. It agrees with the above-mentioned whisker-initiation behaviors in the samples. We thus conclude that in the CUFE sample, a whisker initiates either from a surface grain immediately on top of a GB or from surface grains located on both sides of the same GB. To confirm this conclusion, the correlation between the tin-diffusion sites and whisker formation sites was investigated. Simulation of atom diffusion by molecular dynamics indicated that the dominant tin-diffusion site is a GB when compressive stress is applied in the direction normal to the GB. Investigation of the correlation between the whisker roots and coating microstructures of the CUFE sample showed that the whisker roots were located on top of GB intersections in the coating. These results indicate that whisker-initiation sites are correlated with dominant tin-diffusion sites and that each whisker initiates either from a surface grain located immediately on top of a GB or from surface grains located on both sides of the same GB.

Fine line circuit manufacturing technology with electroless copper plating

Two types of additive processes for fine circuit pattern manufacturing technology using electroless copper plating have been developed. The processes offer high dimensional accuracy. Technical aspects of the additive processes, materials for the fabrication of additive circuits, and the performance of these circuits are reported here.<<ETX>>