In honour of Prof. D r . D r . h . c . P. GORLICH's 70th birthday The gold doping of silicon i s in common use in semiconductor technology for the control of the lifetime of minority carriers. This causes the necessity of searching for and investigating the possible gold doping methods. One of these is the thermal gold diffusion (1, 2). Recently more data become available, indicating the effect of dislocations, grain boundaries etc. on the gold thermal diffusion rate ( 3 to 5). Lambert (3), for example, has obtained the relation kt 0 . 5 (C / c ' ) = (1e , s s I S Swhere C i s the gold concentration incorporated in silicon, C the maximum gold solubility in silicon at this temperature, k the rate factor, and t the diffusion time.The value of k was found to depend on the grow-in dislocation density. It is quite natural since gold migrates in silicon via interstitials and is captured recurrently by vacancies, which a r e injected by dislocations moving during the diffusion annealing.Kastner and Hesse (5) have observed that dislocations within the density range 8 -2 up to 10 cm If silicon samples with introduced dislocations have been annealed at 500 C an increase of the gold solubility with dislocation density was observed. The authors of (5) explained this effect by assuming a production of dislocations complexes, which inject excess vacancies during the diffusion process.introduced by plastic deformation did not influence the gold solubility. 0 It is known that gold in silicon can be introduced by low-temperature (77 to 400 OK) radiation enhanced diffusion (RED) (6, 7). The RED dependences were investigated in (6). In this connection it is interesting to investigate the effect of crys-