Volume 38, Issue 2, 01 February 1967
Index of content:
38(1967); http://dx.doi.org/10.1063/1.1709354View Description Hide Description
The thermal conductivityK of initially n‐type silicon and germanium irradiated at about 30°C at four fast‐neutron doses (1.1×1017, 2.5×1017, 1.7×1018, and 3.4×1018 n/cm2) was measured between 5°K and 300°K. The thermal‐conductivity behaviors of the two irradiated materials differ by significant features: Pronounced shifts of the maximum of K to higher temperatures are observed in germanium after irradiations, leading to a depression of K which is more pronounced below than above the maximum. No noticeable shift, however, appears in silicon after bombardment, the depressions of K on both sides of the peak remaining comparable. The additive thermal resistivity (K −1−K 0 −1) at 20°K is found to increase with the integrated flux φ, approximately as 7.9×10−13 φ0.65 cm·deg/W, and 3.8×10−12 φ0.65 cm·deg/W, for irradiated silicon and germanium, respectively. Similar flux dependences were observed by Vook for 2‐MeV electron‐irradiated silicon and germanium. Using the Callaway model, it is shown that the experimental data for irradiated silicon can be accounted for by considering only an increase in the τ−1=A′ω4 scattering, the factor A′ increasing linearly with the flux. The additional scattering mechanism is most likely to be associated with strain fields arising from bombardment‐induced defects instead of an electron‐phonon interaction, unless one assumes for the latter a Rayleigh‐type scattering law. In agreement with Vook's results on electron‐irradiated germanium, the dominant additional scattering in the present irradiated germanium, at least for the lower doses, appears to be due to electron‐phonon interaction. On the basis of the Keyes model, the difference between bombardment‐induced scatterings in silicon and germanium is attributed to the predominant effects of irradiation in these materials,silicon approaching an intrinsic behavior and germanium becoming highly p type.
38(1967); http://dx.doi.org/10.1063/1.1709355View Description Hide Description
It has been found that the activation field for 180° domain wall motion in a BaTiO3 crystal is increased when the wall intersects an ``a'' domain for the case where the wall is parallel to the edge of the ``a'' domain. It is shown that this increase is only an apparent increase due to a reduction of the mean fieldE 2 acting on the wall in the ``a'' domain region compared with the field E 1 acting on the walls which do not intersect ``a'' domains. If d 1=crystal thickness and d 2=``a'' domain thickness, then E 1/E 2 = (d 1+d 2)/d 1. As a result, it is calculated that the contribution of any ``a'' domain spikes in the surface to the increased surface‐layer activation field is negligible. It is also shown that the motion of 180° walls is determined by the fields acting throughout the whole of the crystal and not just by the fields at the surface of the crystal.
38(1967); http://dx.doi.org/10.1063/1.1709356View Description Hide Description
The elastic constants of NaCl, KCl, and KBr have been measured from room temperature to temperatures near their melting points. A phase‐comparison technique was used to measure the velocity of high‐frequency sound waves in the crystals from which the elastic constants were derived. The elastic constants obtained were found to agree, within experimental error, with previously measured values. The theoretical values of the temperature dependence calculated from the expressions of Mitskevich agree very well with the experimentally obtained values. The value and temperature dependence at high temperatures of the elastic constants (c 11–c 12)/2 are compared to the relative meltingtemperatures of the three crystals. In support of the theory of Thompson, the rapid decrease of this elastic constant can be correlated with the relative temperature of melting.
Temperature Dependence of the Elastic Constants of the Alkali Halides. II. The Solid Solution KCl–KBr38(1967); http://dx.doi.org/10.1063/1.1709357View Description Hide Description
The elastic constants for the solid solution KCl–KBr have been measured from room temperature to 400°C. The general trend of the elastic constants versus composition is in agreement with theory but indicates that the dependence on interionic distance is greater than predicted theoretically. The temperature coefficients of the elastic constants are found to agree fairly well with the theory of Mitskevich for c 11 and ½(c 11–c 12), but those for c 44 were best fit by an expression taken from Leibfried and Hahn. The dependence on composition of ½(c 11–c 12) at high temperature strongly supports the theory of Thompson that melting is due to the instability of the acoustic mode associated with this shear elastic constant.
38(1967); http://dx.doi.org/10.1063/1.1709358View Description Hide Description
The elastic constants of CsCl, CsBr, and CsI have been measured from room temperature to 450°C. Measurements on CsCl were made up to its transition temperature at 473°C but no anomalous variations were found which would differentiate its behavior from that of CsBr and CsI. The results of the measurements are shown to be in agreement with previous experimental work. It is found that the type of temperature variation exhibited by c 44 in the CsCl‐type lattice is very similar to that observed for ½(c 11–c 12) in the NaCl‐type structure. From a consideration of the respective ionic arrangements, these elastic constants can be seen to describe shear mechanisms which could lead to instability with a static expansion of the lattice. From such a treatment, a possible mechanism is suggested to describe the CsCl‐type to NaCl‐type structure transition. This mechanism is shown to be compatible with the existing data.
38(1967); http://dx.doi.org/10.1063/1.1709359View Description Hide Description
The temperature transients arising from instantaneous plane heat sources in a linear heat conductor are analyzed. The thermal conductivity and the specific heat of the conductor are assumed to be temperature independent. Two thermometers are fastened to the conductor. The thermometers possess heat capacity causing inert response. It is found that the position and strength of a heat source can be calculated from the first and second time integral of the observed transient temperature difference between the two thermometers. If the inertia parameter is large, the thermometerheat capacity can be derived from the response to a uniform short heat pulse. The method can be applied to problems of low‐temperature plastic deformation and superconductivity. An experimental example is discussed, in which instantaneous plane heat sources are generated by localized discontinuous tensile deformation at 4.2°K. The positions of the slipped zones and the amount of heat released during each slip are determined.
Correction for Finite‐Pulse‐Time Effects in Very Thin Samples using the Flash Method of Measuring Thermal Diffusivity38(1967); http://dx.doi.org/10.1063/1.1709360View Description Hide Description
A method is described for eliminating errors due to the finite‐pulse‐time effect in the flash technique for measuringthermal diffusivity. The method depends on the employment of a mathematical model which incorporates an empirical function closely describing the actual waveform of the heat pulse produced by a flash lamp. The analytical solution obtained forms the basis of a digital computer data‐reduction procedure devised for calculating the thermal diffusivity from the experimental measurements. The method of data reduction employed leads to an improvement in the precision of measurement heretofore unattainable by the flash technique. This is accomplished, in effect, by replication of the experiment through utilization of the information contained over a wide range of the temperature transient, in addition to that provided by the usual measurement of the time at which the transient reaches half‐maximum.
The analysis described makes it possible to use the flash technique with samples having very small thickness‐to‐diameter ratios. This has the advantage of widening the range of temperatures within which heat losses may be neglected without the introduction of serious error. Experimentally, the method has been successfully tested for some very thin disk‐shaped samples having axial dimensions of the order of (ατ)1/2, where α is the thermal diffusivity and τ is the duration of the pulse.
38(1967); http://dx.doi.org/10.1063/1.1709361View Description Hide Description
A Pochettino‐type rheometer is described which is suitable for polymer solutions or diluted networks, i.e., gels. It directly measures sinusoidally varying stress and strain together with the phase angle between them. The design makes use of the piezoelectric properties of titanateceramics for both the ``driver'' and the ``pickup.'' Thanks to these, the rheometer is rather compact and temperature control is greatly simplified. The sample is located in a narrow gap (0.3 mm) between two coaxial cylinders. The very small displacements (0.1 μ) are measured with a high‐resolution inductive displacement transducer. An important feature of the design is a ``compensator'' ceramic, between the driver and the pickup, which eliminates secondary modes of motion generated by the driver. The rheometer is calibrated by an absolute method which avoids the use of standard viscosity samples. The instrument can measure moduli from 500 to 6×107 N/m2. The frequency range—not yet fully exploited—is from 10 to 3500 cps. Measurements have been made between −80° and +30°C. The temperature range possible with the present design is −85° to +100°C. With the above rheometer, the relaxation spectrum of a 25.5% solution of polymethyl acrylate (PMA) in toluene was obtained. Below −25°C this solution exhibits a phase separation on a microscopic scale, leading to a very broad distribution of relaxation times and analogous to that observed for some heterogeneous polymer systems.
38(1967); http://dx.doi.org/10.1063/1.1709362View Description Hide Description
Dislocation studies in Bi2Te3 were made by etch‐pit technique. All work was done on c planes. The etch marks the dislocations in the c plane by producing etch grooves, and marks the sites of emergence of dislocations from the surface with pyramidal etch pits. Evidence was found for the presence of networks of dislocations lying predominantly in the c plane. Experimental evidence was also found for bending of the dislocations within the crystal being responsible for different size pyramidal etch pits, and flat bottom pits observed on the etched surfaces. The same mechanism was found to be responsible for bringing out new etch pits on longer etching. The slip traces on the c plane were found by controlled surface‐damage studies to be parallel to the binary axes.
38(1967); http://dx.doi.org/10.1063/1.1709363View Description Hide Description
A comprehensive experimental and theoreticalanalysis is presented for the electron‐rich, unignited mode of a cesium‐vapor thermionic diode. The theoretical inverse spacing dependence of the apparent saturation current is found to be in agreement with experiment. The theoretical and experimental coefficient of this dependence are found to be in agreement if one assigns a value Pc =700 cm−1 Torr−1 to the electron‐neutral collision probability. This value is consistent with other available information. Potential distributions are calculated. Full theoretical and experimental volt‐ampere curves are compared. Some of the assumptions in the theoreticalanalysis are examined in detail.
38(1967); http://dx.doi.org/10.1063/1.1709364View Description Hide Description
The interaction of a domain wall with a crystalline defect is frequently represented by means of a potential‐energy function which gives the energy of the wall as a function of position. However, it was shown ten years ago by Rodbell and Bean that the experimental dependence of wall velocity on applied magnetic field is inconsistent with the predictions of a theory in which the interaction of a wall with a defect is taken to be conservative. In the present paper a model is proposed in which this interaction is mainly nonconservative. It is shown that this model is in agreement with wall‐velocity measurements. It also predicts the reversible spikes found on the leading and trailing edges of the switching voltage when a pulse of magnetic field is applied to a square‐loop sample. Finally, it is shown that the model predicts a Rayleigh hysteresis loop. A one‐parameter expression is given for the Rayleigh loop of a single‐crystal window frame.
38(1967); http://dx.doi.org/10.1063/1.1709365View Description Hide Description
Electromagnetic signals have been found to originate from the wave fronts in detonating high explosives. The effect is that of small positive electric charge on the detonation front. For the liquid explosive, nitromethane, the amplitude of the detected signal increases smoothly as the wave approaches the detector. For particulate explosives there is superimposed on the signal a complex oscillatory pattern, the amplitude and frequency distribution of which are dependent on the kind of explosive, the particle size, and the presence of macroscopic disturbances of the detonation wave. Signal detection is achieved with an external passive probe, which has no effect on the explosive system under investigation and requires only a high‐speed oscilloscope to record detonation events. These signals form the basis of a powerful new method for studying propagation times, initiation phenomena, and the structure of the detonation front.
38(1967); http://dx.doi.org/10.1063/1.1709366View Description Hide Description
We have constructed a generator of dynamic Newtonian gravitational‐force‐gradient fields and used it as a signal generator to calibrate the response of the gravitational‐gradient detectors being developed in our research work on gravitational‐mass sensors. The gravitational‐gradient‐field generator is a flat aluminum cylinder 14 cm in diameter, with four holes than can be filled with slugs of different density to create a rotating mass‐quadrupole moment. The generator is mounted on an air‐bearing‐supported motor and rotated at a nominal speed of 44 rps (2640 rpm). Because of the bisymmetric mass distribution, the dynamic gravitational‐gradient fields generated have a frequency of 88 Hz, or twice the rotation frequency. The detector is a 12‐cm‐diam cruciform‐shaped structure which responds to 88 Hz gravitational‐gradient forces. The small (10−11 cm) motions induced in the detector arms are sensed by piezoelectric strain transducers attached to the arms near the point of maximum strain. A simple vacuum system, an iron shield plate, and spring mounts suffice for acoustic and magnetic isolation, since most of the nongravitational noises were generated at 44 Hz, the rotation frequency, rather than at 88 Hz, the gravitational‐gradient frequency. Data taken with four different mass distributions varying from 0 to 1000 g and separation distances varying from 4.8 to 12 cm agree well with the theory, indicating that only gravitational energy was being transmitted from the generator to the detector. The minimum dynamic gravitational‐gradient field observed during this test was 6×10−9 sec−2 or 0.002 of the earth's gradient. The equivalent differential acceleration exerted on the sensor arms by this field was 3×10−11 g's.
38(1967); http://dx.doi.org/10.1063/1.1709367View Description Hide Description
Single‐crystal lithium niobate pumped with pulsed ruby‐laser radiation has been used to convert 1.7‐μ radiation to green light with more than 1% efficiency. A narrow infrared bandwidth of 17 Å, set by the phase‐matching requirement only, allows the up‐converter and photomultiplier to operate in place of a monochromator and infrared detector, and the emission spectrum of a mercury lamp has been thus examined in the region of 1.7 μ. A close agreement between theory and practice has been found in all respects except noise performance. Further studies of this aspect are required.
38(1967); http://dx.doi.org/10.1063/1.1709368View Description Hide Description
The dependence of the critical transport current density on conductor size has been measured for the superconductor, Ti (22 at.% Nb), in transverse fields up to 93 kG. For samples with as similar metallurgical histories as possible, it is found that the critical current is proportional to the cross‐sectional area for samples 0.0024 in. to 0.0501 in. in diameter. More important, the field‐dependence found for critical current densities is given by,where B is the applied transverse magnetic field.
38(1967); http://dx.doi.org/10.1063/1.1709369View Description Hide Description
The results of a microwave investigation of gaseous plasmas produced by the radiation field within the core of a water‐cooled nuclear reactor are reported. It is shown that the plasmas thus formed are applicable in reactor instrumentation and are quite suitable for investigating basic plasma phenomena such as various collision cross sections and electron‐loss processes. Radiometric electron‐temperature measurements indicate that, at adequately high gas pressures (≥50 Torr) in the light noble gases used, the electron gas temperature is at or near the host gas temperature and is independent of the reactor power. The dependence of the electron temperature on the gas pressure has been utilized to determine the electron—neutral collision frequency in neon at different electron energies and in helium at room temperature. Preliminary measurements give a value of 2×10−9 cm3·sec−1 for the recombination coefficient in helium at room temperature.
38(1967); http://dx.doi.org/10.1063/1.1709370View Description Hide Description
This is a mathematical analysis of the conduction‐diffusion equation, which governs the carrier motions in semiconductor materials. The model used for the junctiondevices is based on the Fletcher—Harrick injection relations at junctions and on the vanRoosbroeck conduction—diffusion equation in bulk regions. The exact solution of this analysis is obtained by numerical computation using the Automatic Taylor Series (ATS) method. Besides the numerical solution, there is a closed‐form exact solution for zero recombination; this is the Gunn solution.
The limitations of this analysis are those imposed on the Fletcher—Harrick and vanRoosbroeck developments. They are the abrupt‐junction condition, and the quasineutrality condition. Satisfaction of the first condition depends on devicefabrication. This analysis does not include the effects found in diffused junctions. Satisfaction of the quasineutrality condition depends on the internal electric fields in the devices. This analysis examines these electric fields and keeps watch over its own validity.
The characteristics of semiconductor junctiondevices are analyzed by the conduction—diffusion theory with mathematical uniformity and exactitude. Application of this theory to the p—n and n—n +junctions yields these results. The characteristics of p—njunctions analyzed here basically agree with the results of previous researchers. The n—n +junction exhibits a current‐saturation effect in the forward bias direction. This is due to a limitation on the injected carrier concentration in this device. The p—n—n + and n + —n—n +junctions are analyzed to eliminate metal contacts to the base region, because the quasineutrality condition is violated near these contacts. The three‐layer devices have basically the same characteristics as the two‐layer diodes. In p—i—njunctiondevices with long base regions, the quasineutrality condition is violated in the middle of the base region. In such cases, the solutions of Lampert and Rose and of Baron are more appropriate.
38(1967); http://dx.doi.org/10.1063/1.1709371View Description Hide Description
By comparing field‐ion micrographs of ordered equiatomic Pt–Co alloy with the surface structure expected from a hemisphere of the established L1 0 lattice, it is found that the Co atoms are not imaged. This greatly facilitates the domain structure analysis. Using the FIM, 90‐deg oriented domains and antiphase domains are observed at the atomic scale. It is found that, in general, randomly distributed 90‐deg oriented domains and antiphase domains of size about 20–300 Å constitute the ordered Pt–Co crystal. Additionally, platelets of thickness 20–80 Å, oriented at 90‐deg with respect to the main matrix have been observed. By comparing the ion micrographs of in situannealed specimens at increasing ordering times, it is concluded that both nucleation and homogeneous rearrangement are responsible for the ordering mechanism at low temperature.
38(1967); http://dx.doi.org/10.1063/1.1709372View Description Hide Description
The platinumdeposits overgrown on iridium and tungsten are polycrystalline with a crystal size of 50–500 Å. The orientation is mainly . Some deposits on iridiumgrow epitaxially. A variety of defects, such as twinning, vacancy clusters, and high dislocation density (1011/cm2), are observed. Additionally, some experiments with copper deposits on iridium and tungsten are described. The corrosion of field‐evaporated iridium tips both in air and in the plating solution is studied prior to the application of the electroplating.
38(1967); http://dx.doi.org/10.1063/1.1709373View Description Hide Description
The attenuation of a strong plane shock produced in a target by the impact of a thin striker is studied by the numerical method of characteristics. The calculated results show that late‐stage equivalence exists for impacts in aluminum,copper, and ideal gases with various ratios of specific heats, γ. The shock fronts produced by different impacts approach each other in position, and in strength, at late times provided that du 0 α is constant, where d is the thickness of the striker, u 0 the impact velocity, and α a constant with values of 1.28, 1.50, 1.62, 1.50, and 1.32, for aluminum,copper, and ideal gases with γ equal to 2.0, 1.4, and 1.1, respectively.