Nanostructured devices have the potential to serve as the basis for next-generation energy systems that make use of densely packed interfaces and thin films. One approach to making such devices is to build multilayer structures of large area inside the open volume of a nanostructured template. Here, we report the use of atomic layer deposition to fabricate arrays of metal-insulator-metal nanocapacitors in anodic aluminium oxide nanopores. These highly regular arrays have a capacitance per unit planar area of approximately 10 microF cm-2 for 1-microm-thick anodic aluminium oxide and approximately 100 microF cm-2 for 10-microm-thick anodic aluminium oxide, significantly exceeding previously reported values for metal-insulator-metal capacitors in porous templates. It should be possible to scale devices fabricated with this approach to make viable energy storage systems that provide both high energy density and high power density.

Nanotubes are fabricated by atomic layer deposition (ALD) into nanopore arrays created by anodic aluminum oxide (AAO). A transmission electron microscopy (TEM) methodology is developed and applied to quantify the ALD conformality in the nanopores (thickness as a function of depth), and the results are compared to existing models for ALD conformality. ALD HfO2 nanotubes formed in AAO templates are released by dissolution of the Al2O3, transferred to a grid, and imaged by TEM. An algorithm is devised to automate the quantification of nanotube wall thickness as a function of position along the central axis of the nanotube, by using a cylindrical model for the nanotube. Diffusion-limited depletion occurs in the lower portion of the nanotubes and is characterized by a linear slope of decreasing thickness. Experimentally recorded slopes match well with two simple models of ALD within nanopores presented in the literature. The TEM analysis technique provides a method for the rapid analysis of such nanostructures in general, and is also a means to efficiently quantify ALD profiles in nanostructures for a variety of nanodevice applications.

A recently reported ruthenium molecule, bis(2,6,6-trimethyl-cyclohexadienyl)ruthenium, has been developed and characterized as a precursor for atomic layer deposition (ALD) of ruthenium. This molecule, which has never been reported as an ALD precursor, was developed to address low growth rates, high nucleation barriers, and undesirable precursor phases commonly associated with other Ru precursors such as RuCp and Ru(EtCp) 2 . The newly developed precursor has similar vapor pressure to both RuCp and Ru(EtCp) 2 but offers significant improvement in stability as evaluated by thermogravimetric analysis and differential scanning calorimetry. In an ALD process, it provides good self-limiting growth, with a 0.5 Å/cycle growth rate under saturated dose conditions in a temperature between 250 and 300 °C. Furthermore, the precursor exhibits considerably better nucleation characteristics on SiO 2 , TiO 2 , and H-terminated Si surfaces, compared to RuCp 2 and Ru(EtCp) 2 .

In-situ sensing using mass spectrometry and its use for run-to-run control on a W-CVD cluster tool AIP Conf.Process diagnostics and thickness metrology using in situ mass spectrometry for the chemical vapor deposition of W from H 2 / WF 6 Real-time, in situ chemical sensing has been applied to achieve reaction metrology and advanced process control in a low pressure tungsten chemical vapor deposition process based on WF 6 and SiH 4 reactants ͑silane reduction process͒. Using mass spectrometry as the sensor to detect both product generation (H 2 ) and reactant depletion (SiH 4 ) at wafer temperature of 200-250°C, these signals provided a direct real-time measurement of deposited film thickness with an uncertainty less than 2%, and this thickness metrology signal was employed to achieve real-time process end point control. When reactant conversion rates are sufficient ͑ϳ20% in this case͒ as often occurs in manufacturing processes, the thickness metrology ͑1.0%-1.5%͒ and control ͑ϳ1.5%-2.0%͒ accuracies are in the regime needed for meaningful application of advanced process control. Since the in situ sensor delivers a metrology signal in real time, real-time process control is achieved, enabling compensation for random process disturbances during an individual process cycle as well as for systematic wafer-to-wafer process drifts. These results are promising for manufacturing from the standpoints of metrology accuracy and application in real-time control.

In situ quadrupole mass spectrometry (QMS) has been integrated to an atomic layer deposition (ALD) reactor to achieve real-time chemical diagnostic and wafer-state metrology. The process investigated was tungsten ALD using WF6 and SiH4. The UHV-based substrate-heated ALD reactor incorporated a minireactor chamber to simulate the small reaction volume anticipated for manufacturing tools in order to achieve adequate throughput. Mass spectrometry revealed essential surface reaction dynamics through real-time signals associated with by-product generation as well as reactant introduction and depletion for each ALD half-cycle. The by-product QMS signal was then integrated in real time over each exposure and plotted against process cycle number to directly observe ALD film growth, leading to two valuable metrologies. First, the integrated by-product QMS value changes with cycle number, directly reflecting the nucleation kinetics. Specifically, QMS values increase with cycle number during the nucleation phase and then saturates as the film growth enters its steady-state growth phase. Second, summing the integrated by-product QMS signals over an entire deposition run provides an immediate measure of film thickness. The growth kinetics as measured by QMS is consistent with ex situ film characterization and is strongly dependent on process conditions and reactor chamber status. In the latter case, a clear first wafer effect was apparent when the system was left idle for a few hours, resulting in an apparent QMS signal difference during nucleation phase between the first wafer and nonfirst wafer cases. The dependence of QMS signals on chamber status is attributed to parallel reactions on the chamber wall, where different gas exposure history is encountered. The first wafer effect can be explained in a quantitative manner by considering the chamber wall as an additional wafer inside the ALD reactor. The first wafer effects can be reduced by proper preprocess treatment, and the linear correlation between QMS measurement and film thickness suggests a promising start for QMS-based ALD film thickness metrology.