A one-dimensional heat conduction model is developed for a phase change random access memory device with an 8F2 memory cell structure (F=0.15 μm). The required current level for a reset operation, which corresponds to the phase switching from a crystalline (“1” state) to an amorphous phase (“0” state) of Ge2Sb2Te5, was investigated by calculating one-dimensional temperature profiles for the memory cell structure. It is revealed that a reset operation is not achieved at the current level (2 mA) reported for existing devices with a subquarter micron plug size when only TiN is used as a resistive heater. However, it is possible when an additional heating layer of 5 nm thickness is inserted between the TiN and Ge2Sb2Te5 layers, for which the electrical resistivity ρelec is higher than 105 μΩ cm, and the thermal conductivity κ and specific heat c are as low as those of Ge2Sb2Te5. In addition, it is shown that a reset operation at a low current level of 1 mA can be realized in this memory cell when amorphous carbon (κ=0.2 W/m K and ρelec=106 μΩ cm) is used as an additional heating layer. It is believed that this relatively simple one-dimensional heat conduction model is a useful tool for analyzing the device operation of phase change random access memory devices and for selecting the proper conditions for an additional heating layer allowing for low-current operation.

The above results suggest that hedgehog signaling is involved in the differentiation of gastric progenitor cells during the gastric ulcer repair process.

The properties of WNxCy films deposited by atomic layer deposition (ALD) using WF6, NH3, and triethyl boron as source gases were characterized as a diffusion barrier for copper metallization. It is noted that the as-deposited film shows an extremely low resistivity of about 350 μΩ cm with a film density of 15.37 g/cm3. The film composition measured from Rutherford backscattering spectrometry shows W, C, and N of ∼48, 32, and 20 at. %, respectively. Transmission electron microscopy analyses show that the as-deposited film is composed of face-centered-cubic phase with a lattice parameter similar to both β-WC1−x and β-W2N with an equiaxed microstructure. The barrier property of this ALD–WNxCy film at a nominal thickness of 12 nm deposited between Cu and Si fails only after annealing at 700 °C for 30 min.

Ruthenium (Ru) thin films were grown on thermally-grown SiO 2 substrates by thermal atomic layer deposition (ALD) using a sequential supply of a zero metal valence precursor, isopropyl-methylbenzene-cyclohexadiene Ru(0) (IMBCHRu, C 16 H 22 Ru) and molecular oxygen (O 2 ) at substrate temperatures ranging from 185 to 310 C. The growth rate at 185 C was approximately 0.059 nm/cycle but its resistivity was >3000 X cm. When the deposition was done at 200 C, the resistivity was decreased drastically to $100 X cm, and the film growth rate increased to 0.075 nm/cycle. An ALD temperature window from 225 to 270 C was observed. A high growth rate of 0.086-0.089 nm/cycle was obtained at this ALD temperature window. The film deposited at 270 C showed a minimum resistivity of $30 X cm and a high density of 11.7 g/cm 3 and with no impurities in the film, such as oxygen and carbon. At 310 C, the growth rate increased to 0.136 nm/cycle due to a partial decomposition of the precursor. In addition, the film resistivity increased slightly to $40 X cm with the incorporation of carbon and the formation of a less-dense film. The step coverage of the ALD-Ru film was dependent on the dimensions of the contact and deposition temperature. At the contact with an aspect ratio of $4.6 (top opening diameter: 80 nm), the step coverage was excellent irrespective of the deposition temperature. However, at the contact with an aspect ratio of $25, the step coverage of the film deposited at 310 C (or above ALD temperature window) was degraded, even though those prepared within ALD temperature window were $100%. Finally, ALD-Ru film was used successfully as a seed layer for Cu electroplating.

Ruthenium (Ru)-based ternary thin films (RuAlO) were prepared by thermal atomic layer deposition (ALD) with repeated supercycles consisting of Ru and Al 2 O 3 ALD sub-cycles at 225 C. The step coverage of ALD-RuAlO was excellent, around 93% at contact holes with an aspect ratio of $29 (top-opening diameter: $74 nm). Transmission electron microscopy analysis showed that RuAlO films formed with non-columnar grains and a nano-crystalline microstructure consisting of Ru nano-crystals separated by amorphous Al 2 O 3 . The sheet resistance and X-ray diffraction showed that the structure of Cu (100 nm)/RuAlO (15 nm)/Si was stable after annealing at 650 C for 30 min. Fifty nanometer-thick Cu was electrodeposited directly on RuAlO film, suggesting that it could be a viable candidate as a Cu direct plateable diffusion barrier.The current structure of Cu interconnects consists of electroplated (EP)-Cu that is responsible for the most of the current and an underlying stack of a relatively high-resistive Ta/TaN diffusion barrier and Cu seed layer for the Cu EP process, which is deposited mostly by physical vapor deposition (PVD). 1 However, with the continuous scaling-down of devices, the filling of EP-Cu into patterned features becomes increasingly difficult, which is aggravated by the underlying thick diffusion barrier and seed layer. Moreover, an unwanted and drastic increase in Cu wire resistance occurs due to the size effect on the resistivity of metal films. 2,3 Both problems can be addressed by increasing the volume of EP-Cu filled in the patterned features by the direct plating of Cu because the EP of Cu can be achieved on a diffusion barrier without a seed layer. The conformality, thickness controllability, and large-area uniformity of the process for a Cu direct-plateable diffusion barrier need to be considered due to continuous scaling of the devices. Atomic layer deposition (ALD) is a viable solution for depositing a Cu directplateable diffusion barrier because it employs a self-limiting film growth mode through surface-saturated reactions, which enables atomic-scale control of the film thickness with excellent step coverage. 4 Ru has been suggested as a diffusion barrier for seedless Cu interconnects. 5,6 In addition, Ru films have been grown successfully by ALD using a variety of Ru metallorganic precursors with excellent step coverage. 7-12 However, Ru itself is not a suitable diffusion barrier against Cu due mainly to its poor microstructure with polycrystalline columnar grains. 13,14 Indeed, many studies demonstrated that the microstructure plays an important role in the resulting diffusion barrier performance. 15 Two types of approaches were reported to address the poor diffusion barrier performance and obtain reliable seedless Cu interconnects with Ru. The first is to use Ru as a seed layer and combine it with superior materials to function as a diffusion barrier against Cu. Sputtered-deposited Ru/ TaN, 16 Ru/Ta, 17 Ru/WN x 14 bilayers, and atomic layer deposited Ru/TaCN bilayer 18 have been suggested and t...