A. Reisman

Chemical Vapor Deposition of Copper from Copper (II) Hexafluoroacetylacetonate

Thermally activated decomposition of the vapor phase of copper (II) hexafluoroacetylacetonate was studied

The modelling of Silicon oxidation from 1 x 10 –5 to 20 atmospheres

Studies of (100) silicon oxidation in dry oxygen at 800° C and 20 atmospheres showed the data to be modelled accurately over a wide thickness range by the simple equationx =a(tg +to)b, wherex is the final measured thickness,a andb are constants,tg is the time for growth measured in an oxidation experiment, andto is the time to grow an oxide of thickness,xo, already present on the silicon surface, and/or formed as a consequence of furnace ramp-up and ramp-down sequences. The values forto, andxo are determined readily fromx-t data. Employing this power law model, there is no evidence of an “anomalous” oxidation region at small thicknesses, or of a linear oxidation region in the earlier stages of oxidation. These surprising results led to an analysis of most of the published silicon “dry” oxidation data from 1965 to the present time to determine whether this power law model was general, or representative only of the data obtained in the present study. It was found that all of the published data, including that published by Deal and Grove in 1965, could be mathematically modelled precisely and simply in all of the thickness ranges reported, and at partial pressures of oxygen from less than 1 × 10−5 up to 20 atm. The critical parameterb was found to range from less than 0.25 to about 0.8, the value decreasing with decreasing partial or total pressure of oxygen. The constanta represents the value of oxide thickness at unit time, and for constant values ofb andto, causes a shift of thex-t curves along the thickness axis. The physical origin(s) for the model is not known. However, as the data appear to suggest a diffusion limited oxidation is in play over the entire thickness regime, it is likely that the diffusion coefficient for the process varies with oxidation time, therefore oxide thickness.

Thermal stress analysis of a multichip package design

The authors present a thermal analysis of a thin-film multichip package design, with emphasis on thermally induced stress in the critical package components. The package uses flip-chip solder bonding and thin-film interconnections between chips. Indium was chosen as the die attachment medium between each chip and the water-cooled heat sink. A methodology is given to estimate the stresses in the structure during a power-up. Finite-difference and finite-element computer simulations were used to calculate the temperature and stress distributions under both transient and steady-state conditions. It is shown how thermal gradients, expansion mismatches, and global bending of the structure determine the stress distribution. The components in the module have various thermal time constants, and the stresses during a transient are related to the rate at which each component heats up. For instance, the chips and the heat sink complete 70% of their temperature rise in the first 200 ms, but the substrate takes over 10 s to reach 70% of its steady-state temperature rise. Therefore, even if a design is optimized to be thermal expansion matched under operating conditions, stresses can develop during a transient. >

Defect generation in silicon dioxide from soft x‐ray synchrotron radiation

Generation of fixed positive charge, neutral electron traps, and fixed negative charge in SiO2 due to exposure to monochromatic x‐ray radiation in the photon energy range from 300 to 1000 eV in a synchrotron source is reported upon. At a constant exposure level of 2×107 rads, the number of defects generated is approximately independent of x‐ray photon energy. The generated defects, independent of x‐ray radiation energy, show normal post‐metal annealing characteristics.

Cu CVD from Copper(II) Hexafluoroacetylacetonate I . A Cold Wall Reactor Design, Blanket Growth Rate, and Natural Selectivity

An atmospheric pressure cold wall reactor was designed and built for the purpose of studying the thermal decomposition of Cu(hfa) 2 as a repair technique for broken copper interconnection lines, using thermally biased substrates, and a laser to heat localized areas to a temperature appropriate for the deposition of pure copper. In this paper, a discussion of the design is presented, and theoretical and experimental blanket copper deposition growth rates in argon and forming gas are discussed. The primary goals of the present work were the design of the system, the determination of blanket growth rate characteristics, and examination of the tendency for natural selectivity to occur on two contiguous dielectric materials, SiO 2 and polyimide. In studies of natural selectivity, deposition was observed on SiO 2 , but not on contiguous polyimide at substrate temperatures of 340°C down to 270°C or lower, using argon as a carrier/diluent, and from 270°C down to 150°C, using argon/hydrogen mixtures (9 :1) as a carrier/diluent.

Cu CVD from Copper(II) Hexafluoroacetylacetonate II . Laser‐Assisted Selective Area Deposition

The laser-assisted chemical vapor deposition of copper onto thermal SiO 2 -overcoated Si wafers in a cold wall atmospheric pressure reactor, using Cu(hfa) 2 in Ar/H 2 (10%), or argon as carrier/diluent is discussed. The substrate was biased thermally at 130°C in Ar/H 2 or 200°C in Ar. A multiline (λ= 488 to 514 nm) continuous wave 4 μm focused spot diameter, 150 mW argon ion laser was used to heat the spot to a temperature at, or above, the required decomposition temperature, ≥150°C in Ar/H 2 and ≥250°C in Ar. The steady-state temperature on the SiO 2 prior to the beginning of deposition was estimated using a modified Lax model, since there is no convenient way of measuring it experimentally. Under the conditions employed, a maximum copper spot growth rate of 0.12 μm/min was obtained at 0.095 W laser power. The laser-assisted growth rates in the present system can be predicted within a factor of two from the kinetic model of Ehrlich and Tsao. An incubation time which varies inversely with power, and with the composition of the carrier gas was observed on SiO 2 , due to the low absorption of the incident energy in the SiO 2 at the small laser power used. Auger electron spectroscopy, of the copper deposits has shown small levels of carbon and oxygen. It is possible that the temperature at the center of the beam spot is greater than 400°C after copper begins to nucleate, and that the organic ligands undergo some fragmentation, or that some degree of oxidation and carbonaceous contamination occurs upon exposure to air when specimens are removed from the reactor. Using the described deposition methods, it should be possible to effect repair of broken copper lines on a microelectronic package in a minute or two.

Challenges in advanced semiconductor technology for high-performance and supercomputer applications

This article provides a review of the capabilities, future directions, and technology challenges for semiconductor chips and packages as they apply to high-performance and supercomputer applications. Semiconductor chip technology has resulted in dramatic device density improvements over the last 20 years. Scaling theory predicts that continued improvements will be possible if the technological problems associated with patterning, doping, interconnection, density, yield, and cost can be solved. The issues associated with these challenges are discussed. Finally, the packaging needs to support advanced chip technologies are reviewed.

Studies of the possible reaction of WF6 with SiO2 and Si3N4 at several temperatures

Chemical vapor deposition (CVD) of tungsten by the hydrogen reduction of WF6 is increasingly important in very large scale integration technology. There is, however, a concern about possible interference of this CVD process by the reaction of WF 6 with the SiO 2 mask material. Because of disagreement between thermodynamic calculations and some experimental evidence concerning this reaction, the reaction was examined by several techniques not previously employed. Oxide coated silicon wafers were exposed to the hexafluoride in argon, with and without hydrogen present, at several temperatures from room temperature to 400 o C

Nitridation of silicon in a multiwafer plasma system

Silicon wafers were nitrided in a multiwafer plasma system at low temperatures (< 850°C). An argon plasma (400 kHz rf plasma) was used to which small quantities (approximately 2–8 %) of NH3, N2 or mixtures of N2 and H2 were added. As the rf power was increased, the film thickness as well as the etch rate (in buffered HF) increased. The rate of film growth was found to be slower than that for oxidation in a similar type of plasma system. The effects of variation of power and gas composition on film composition and etch rate are discussed.

A study of X-ray damage effects on the short channel behavior of IGFET's

Ionizing particles and radiation may play an important, albeit undesirable role in the processing of VLSI and ULSI circuits in that they can generate bulk charge in the gate insulator of IGFETs. In this regard, there is conflicting information in the literature on the effects of ionizing radiation on short channel phenomena in IGFETs. For example, Peckeraret al. in 1983 claimed that the effective channel length increases when positive coulombic charge is introduced during irradiation, resulting in a decrease in the short channel effect. Schrankleret al. in 1985 claimed in an experimental study, on the other hand, using 28.0 nm thick gate oxides and 0.9–10 μm channel lengths, that the effect is increased,i.e., the short channel effect begins at longer channel lengths. Wilson and Blue in 1982, in a theoretical study concluded that other than a uniform downward shift in theVT-channel length curve due to the presence of insulator net fixed positive charge, no effect should be observed. Because of these conflicting reports, it was decided to evaluate this behavior using two different background doping levels inn-channel structures, with physical channel lengths ranging between 1.5 and 10 μm, in 0.1 and 0.5 gWcm (100) Si. To further explore the situation, gate oxide (grown at 1000° C in O2 containing 4.5% HC1) thicknesses were varied from 17.0–35.0 nm, and the absorbed radiation dose using Al-Kα (1.5 keV) x-rays was varied between 2.4 × 106 rad (SiO2) and 2.4 × 107 rad (SiO2). For all conditions studied above, a uniform downward shift in the VT-Channel length curve was observed, essentially corroborating the theoretical conclusions of Wilson and Blue. In addition to the above, the effects of intentionally doping the gate insulator with boron (1.2 × 1012 B+ cm−2) implanted at 8 and 10 keV into 25.0 nm and 31.4 nm oxides, respectively, on short channel effects were evaluated for devices grown onp-type 0.5 Ω.cm substrates. Unlike the devices which did not have excess boron intentionally implanted into the gate insulator, it was found that higher concentrations of boron (2.0 × 1017 cm−3 in the insulator via implantation as compared to 4.2 × 1016 cm−3 incorporated in oxides during the oxide growth on 0.5 Ω.cm type (100) Silicon) leads to smaller short channel effects in unirradiated devices. On the other hand, these heavily doped oxides show a distinct worsening of the short channel effect after exposure to 2.4 × 107 rad (SiO2) using Al-Kα radiation. Thus, while normal devices exhibit little if any short channel improvement, or degradation following irradiation, intentionally doped insulators show an improvement in short channel characteristics prior to irradiation, and a worsening of the short channel effect following irradiation.