Hierarchically Structured Materials

We have fabricated arrays of siliconfield emitters using semiconductorlithography techniques. The density of the tips was 105/cm2. The maximum current that can be extracted from each emitter is limited by resistive heating. We have investigated how the electron current emitted changes under constant applied voltage. We found that the current is very sensitive to the vacuum conditions. We attribute this to sputtering of the emitters due to ionized residual gas molecules. The poorer the vacuum, the higher the instability in the current. We studied this phenomenon at 10−6 and 10−8 Torr. The model of two concentric spherical shells is used to obtain the ion energy distribution. This is then used to calculate the rate of ion bombardment and the rate of atoms sputtered. A lifetime of the tip can be deduced from these calculations.  

We have used Raman spectroscopy, large‐ and small‐angle x‐ray diffraction spectroscopy of sputter‐deposited, vacuum‐annealed, soft x‐ray Mo/Si thin‐film multilayers to study the physics of silicide formation. Two sets of multilayer samples with d‐spacing 8.4 and 2.0 nm have been studied. Annealing at temperatures above 800 °C causes a gradual formation of amorphous MoSi2 interfaces between the Si and Mo layers. The transition from amorphous to crystalline MoSi2 is abrupt. The experimental results indicate that nucleation is the dominant process for the early stage and crystallization is the dominant process after nucleation is well advanced. In the thicker multilayer, a portion of the siliconcrystallizes during annealing and a strong Raman signal is observed. An advantage of Raman spectroscopy is that the Raman signal of the silicide is observed even before the presence of MoSi2 can be seen using x‐ray diffraction. This study indicates that Raman spectroscopy is an effective technique for characterizing the formation of crystalline silicides.

Scanning electron microscopy, atomic force microscopy, and Raman spectroscopy were used to characterize the microstructure of photoluminescent porous silicon (PS) layers formed by the anodic etching (HF:H2O:ethanol), at various current densities, of p‐type (100) silicon wafers possessing resistivity in the range 1–2 Ω cm. Existing models for the origin of luminescence in PS are not supported by our observations. Cross‐sectional as well as surface atomic force micrographs show the material to be clumpy rather than columnar; rodlike structures are not observed down to a scale of 40 nm. A three‐dimensional model of the mesostructure of porous silicon is discussed. Room‐temperature Raman scattering measurements show no evidence for a‐Si:H or polysilanes and the material reported here is composed of 10 nm roughly spherical Si nanocrytallites rather than 3 nm wires postulated in standard quantum confinement models.

Our group is studying the structure and interfaces of soft x-ray multilayers by various techniques including x-ray diffraction and Raman spectroscopy. Raman spectroscopy is particularly useful since it is sensitive to the identity of individual bonds and thus can potentially characterize the abruptness of interfaces in multilayers. Blocking interfacial mixing is very important in achieving and maintaining high reflectivity. We report our studies of the as-deposited and postannealed structure of Mo/Si and W/C multilayers. A high normal- incidence, peak reflectance is mandatory for imaging applications that involve many reflections. The reported theoretical and achieved reflectances of the Mo/Si system are 80% and 65%, respectively. This loss of 15% can bring about a six-fold loss in system throughput in the eight-reflection system contemplated. The interfaces in the Mo/Si system are thought to play a significant role in the degrading reflectance so characterization techniques which have interfacial sensitivity are particularly important. The Mo/Si multilayer system is susceptible to Raman characterization since both the a-Si spacer layer and the MoSi2 compound which forms at the interface have Raman active modes. In this paper we report the first Raman studies, to the best of our knowledge, of the a-Si layers and their crystallization and the crystallization of the Mo/Si interface of the multilayer brought about by a one-hour 1000 degree(s)C anneal. These changes are apparent in the Raman spectra before they can be unambiguously detected by x-ray diffraction.

We have produced arrays of 10,000 sharp p-type silicon points using an etch plus oxidation method. These points were used as electron emitters. No high vacuum cesiation or high temperature cleaning was needed to observe the electron emission. These are seen to be photosensitive sources of electrons at 200 K and 300 K. They were also used to produce AlKα x-rays. This constitutes the first use of etched, point arrays for generating electrons for x-ray sources.

We have studied at cryogenic temperatures photoluminescence features which lie more than 0.15 eV below the band edge in ZnxCd1-xTe (0 less-than-or-equal-to x less-than-or-equal-to 0.09) crystals. The same features, namely a defect band which lies at about 0.13-0.20 eV below the band-gap energy and a peak at 1.1 eV, that are observed in pure CdTe samples are observed in these alloy materials. In annealed samples we observe that the 1.1-eV feature, which has been attributed to tellurium vacancies, increases with fast cooling. Increased concentrations of tellurium vacancies can be understood in terms of the phase diagram of CdTe which indicates that higher concentrations of excess Cd appear in CdTe quenched from high temperatures. We also observe an absorption transition near 1.1 eV by photothermal deflection spectroscopy (PTDS). The PTDS phase shifts show that the deep defect is a bulk effect rather than a surface effect. The well-defined absorption peak suggests that the states contributing to the 1.1-eV transition are both localized. Our results also suggest that the defect band which lies 0.13 eV below the band gap (1.48 eV in CdTe) may also be related to tellurium vacancies. However, the fact that the ratio of intensities between this defect band and the 1.1-eV feature is highly variable suggests that the relationship is not simple. The origin of the defect band and its phonon replicas remains controversial.