An investigation was made of tantalum films sputtered in argon containing individual small amounts of nitrogen, methane, and oxygen. With argon alone, 1000-Å bcc tantalum films deposited on glass substrates heated to 400°C had a specific resistivity four times the bulk value, with the increase attributable to a reactive gas residual pressure of 10−6 to 10−5 Torr. The mixed argon-nitrogen experiments produced initially hcp Ta2N, and at higher pressures an fcc phase attributed to a new TaN structure. With methane only fcc TaC was obtained, while with oxygen an amorphous phase began to grow initially, and at higher pressures was dominant with a structure identical to anodically formed Ta2O5. The nitrides and carbide have specific resistivities from 200–300×10−6 Ω cm, and temperature coefficients between +3×10−4 and −2×10−4 deg−1. With increasing oxygen content, the specific resistivity increases approximately exponentially; the positive temperature coefficient decreases, becomes zero before the precipitation of Ta2O5, and then increases negatively at a rapid rate at the appearance of Ta2O5.

No abstract

A general twinning matrix applicable to any crystal system is developed in tensor notation. The symmetry of the cubic system results (1) in the equivalence for twinning of orthogonal planar groups, (2) in successive twinning yielding the same lattice as twinning on certain higher order planes. Selection rules for the calculation of interstitial points in cubic systems show that the generalized reciprocal lattice has a periodicity which is the product of (h) 2 = (hl) z-4-(//2) 2 + (//3) 1 and of the periodicity of the host lattice. Double diffraction produces, in the generalized lattice of all possible relpoints, a large number of points additional to the twin points. In general, the periodicity of noncubic systems is infinite. The hexagonal system, both close-packed and rhombohedral, is treated. Lattice transformations may permit a pseudo-cubic approximation to the generalized lattice.

Contrast and resolution in electron micrographs from thin replica films are determined by the geometrical relationships between the directions of incidence of the condensing atom beam and the local surface normal, during film formation by evaporation in vacuo, and the direction of incidence of the electron beam, during subsequent exposure in the microscope. Replica films may be formed of any material suitable for vacuum evaporation. Metal atoms in general tend to stick where they strike, moving only short distances, 100 Å or less, to nucleating centers where they form small crystallites. Oxides such as silica and silicon monoxide, and also the semi‐metal germanium, form amorphous films. A portion of the incident material, about 50% in the case of silica, migrates large distances, 5000 Å or more, before finally condensing; the remainder sticks where it first strikes the surface. The existence of a minimum perceptible mass thickness difference, about 0.7 μg/cm2 for 50 kv electrons, results in an optimum replica mass thickness of about 10 μg/cm2. The resolution of the replica film is proportional to its linear thickness and hence is inversely proportional to its density. Micrographs of silica, chromium, gold‐manganin, aluminum, aluminum‐platinum‐chromium and germanium replicas are shown. The importance of stereoscopic methods in interpretation of micrographs is discussed.

Electrodeposited uniaxial Permalloy films 10 000 Å thick were annealed for 3 min at temperatures up to 400°C in an easy direction field greater than Hk. Dispersion, α90, increases from 1.8° as plated to about 3.3° with annealing at 325°C. In the same temperature interval, skew and Hk remain unchanged. Above 325°C the magnetic properties change rapidly towards isotropic behavior. The onset of this effect is correlated with a large change in crystallite size. Transmission electron micrographs show a small increase in crystallite size D from 150 Å as-plated to 500 Å for annealing at 325°C, and a rapid size increase thereafter. The data for the change in α90 with D does not agree with theoretical predictions of Hoffmann; experimentally α90 is found to be proportional to D0.5 for annealing to 325°C.