JOURNAL OF MATERIALS SCIENCE 20 (1985)1739-1747 Doping effects on indentation plasticity and fracture in germanium S.G. ROBERTS, P. PIROUZ, P.B. HIRSCH Department of Metallurgy and Science of Materials, University of Oxford, UK Vickers micro-indentation tests have been performed in the temperature range 20 to 420 ~ C on the {0 0 1 } surfaces of germanium crystals of three different dopings: "intrinsic", heavily doped p-type and heavily doped n-type. Indentation sizes, disloc- ation rosette sizes and median/radial crack lengths were measured. Rosette sizes were found to depend strongly on doping, being respectively larger and smaller than in intrinsic material for n-type and p-type specimens, over the temperature range 20 to 420 ~ C. This result correlates well with dislocation velocity measurements in germanium. Inden- tation size (hardness) was found to vary with doping above ~ 300 ~ C, hardness increasing from n-type through intrinsic to p-type material. Crack lengths, as a function of tem- perature, showed a sharp transition (to much shorter crack lengths) at a well-defined tem- perature; this ductile/brittle transition temperature was found to depend on doping, being lowest for n-type (~ 290 ~ C) and highest for p-type (~ 400 ~ C). This is the first observation of a relation between a fracture parameter and bulk electronic doping. 1. Introduction Dislocation velocities in semiconductors have been known for some time to be dependent on the type and concentration of electrically active dopants [1-3]. Several models have been proposed for this effect [4-8] and these have been critically reviewed by Hirsch [9]. Recently, it has been shown that, in silicon, such doping can alter the value of the lower yield stress [10] and affect the extent of dislocation "rosettes" around micro- hardness indentations [ 11]; a model was developed relating the sizes of the rosettes to a minimum stress for motion of the dislocations. The work described here extends these investigations to germanium. This material is of interest for a num- ber of reasons: 1. In silicon, both n- and p-doping were found to increase dislocation velocity with respect to that for intrinsic material. In germanium, n- and p-doping have been shown to have opposite effects on dislocation mobility, the order of increasing mobility being p-doped, intrinsic, n-doped [1, 12]. 2. Germanium is more plastic at low tem- peratures than silicon, allowing experiments 0022-2461/85 $03.00 + .12 to be carried out over a range of temperatures with the low-temperature testing equipment currently available to us. 3. Germanium was found to etch much more easily than silicon, making reliable results more readily obtainable. As in the experiments on silicon [11], measure- ments were made of the hardness values and rosette sizes. Also, large changes in cracking behaviour with temperature were observed, and so the crack lengths were measured as a function of temperature, load and doping. These experiments form part of a wider study aimed at investigating doping effects on the mech- anical behaviour of engineering ceramics such as silicon carbide and diamond. For these materials, large single crystals suitable for direct dislocation velocity measurements or compression testing are not easily available, and so indentation-based tech- niques will be necessary. The tests on silicon and germanium were performed to show the applicability of indentation testing to the investi- gation of the doping effects, and to correlate data thus obtained with data from compression and bend tests (e.g. [1-3, 10, 12]). 1985 Chapman and Hall Ltd. 1739