318 



R. ARNAL AND M. SOREL 



A stacking fault in a cubic face centered crystal is 

 a 1 1 1 -plane which is in a false position with respect 

 to its neighbouring planes. In the case of a stacking 

 fault, the periodicity of the crystal lattice is disturbed 

 across an extended surface. Fig. 3 probably shows 

 such faults. When this type of plane ends inside the 

 crystal, its edge is a so-called "partial dislocation". 



In the present work, the steel was rolled in order 

 to produce dislocations uniformly throughout the 

 whole specimen. The pictures show qualitatively the 

 general elTects of cold working. A detailed evalua- 

 tion in terms of the theory of dislocations has yet to 

 be made. 



Fig. 1 shows how in a slightly rolled specimen 

 ( 1-2 "„ reduction in thickness) dislocation lines begin 

 to spread out from the grain boundary along slip- 

 planes into the interior of the grains. With heavier 

 rolling the density of dislocation increases very fast 

 (fig. 2). In a 25 »„ reduced specimen the slip-planes 

 become curved (fig. 4). 



For a study on dislocations, stainless steel is con- 

 venient especially for the observation of static con- 

 figurations, because the annealing temperature, 

 where the dislocations move, is too high to be 

 attained by heating the object by the electron beam. 

 The situation is different with aluminium, however, 



as it is possible with this metal to study dislocation 

 movements as shown by Whelan, Home, Hirsch (8). 

 So, for this type of research these two materials are 

 in some ways complementary. 



The present work has been sponsored by the Union 

 Miniere du Haul Katanga and the Battelle Memorial 

 Institute. I wish to thank Dr. W. Siegfried for having 

 suggested the research. Our work has been undertaken 

 and the main results have been obtained independently 

 from the group Whelan, Home, Hirsch. I thank Dr. P. B. 

 Hirsch for useful discussion on the interpretation of some 

 special points, and Dr. J. W. Menter for his criticism of 

 the manuscript. 



References 



1. Amelinckx, S., Phil. Mag. 1 (8th Series) 269 (1956). 



2. BiLLiG, E., Internationales Kolloqiiium iiber Halbleiter 



und Phosphore. Garmisch-Partenkirclien, August 1956. 



3. Castaing, R., Rev. met. 52, 669 (1955). 



4. CoTTRELL, A. H., Dislocations and Plastic Flow in Crystals. 



Oxford, 1953. 



5. HEiDFNRriCH, R. D., /. Appl. Phys. 20, 993 (1949). 



6. Read, W. T. Jr., Dislocations in Crystals. New York, 



1953. 



7. Wehner, G. K., Phys. Rev. 102. 690 (1956). 



8. Whelan, M. J., Horne, R. W., and Hirsch, P. B., 



Annual Conference of the Electron Microscopy Group 

 of the Institute of Physics. Reading (England), July 

 1956. 



Migrations of Grain Boundaries Studied with the Emission 



Electron Microscope 



R. Arnal and M. Sorel 



Lahoratoire d'E/eclroniqiie et de Radioelectricite de la Faciilte des Sciences de Paris 



Since the growth of metallic crystals can be ob- 

 served continuously with the help of cinematographic 

 recording, it is possible to get quantitative informa- 

 tion on the cinetics of grain boundaries. 



It was noticed, some time ago, that there was a 

 geometrical analogy between crystal boundaries and 

 two-dimensional soap froth; in this latter case, the 

 mechanism is well known; the equilibrium of inter- 

 faces is due to the presence of surface tensions. 



To show the analogy in a better way, we have 

 tried to observe bidimensional crystals in order to 

 have the same conditions as for soap froth. On the 

 other hand, when the mean diameter of crystals is 

 smaller than the sample thickness, the growth or 

 the disappearance of an internal crystal carries a 

 change in topography and one can only observe 

 secondary effects, the internal cause of which is 

 unknown. 



The emission electron microscope has been used 

 with a small magnification (30 times for a screen 

 distance of 70 cm) and the sample heated by elec- 

 tronic bombardment is about 0.2-0.3 mm thick. 



On a screen (8 8 cm), it is possible to see several 

 crystals the diameter of which is greater than the 

 sample thickness. As soon as the diameter of the 

 image of a crystal becomes greater than j^„ of the side 

 of the screen, one can say that it is the bidimensional 

 case. 



The fluorescent screen is filmed at the rigorously 

 constant rate of two frames per second. During the 

 projection, the crystal growth is therefore shown ten 

 times the speed. The luminous intensity is measured 

 with a photomultiplier, large range variations of 

 luminosity are balanced by a hand-regulated dia- 

 phragm. 



The sample is heated by electronic bombardment 

 and the temperature is measured by a chromel- 

 alumel thermocouple, directly soldered on to the 

 sample. 



After testing many metals, titanium, produced by 

 Pechiney, has proved to be the material for which 

 the crystal growth after the phase transformation 

 (at about 1100 C) is the quickest and the more in- 

 teresting. 



