Thin Solid Films, 92 (1982) L63--L65
L63
Letter
Remarks on gas bubble growth in silver bicrystalline films
R. PAREJA AND M. A. PEDROSA
Departamento de Fisica del Estado Srlido, Facultad de Ciencias Fisicas, Universidad Complutense de
Madrid, Madrid 3 (Spain)
(Received April 20, 1982; accepted May 13, 1982)
In the preceding letter, Lloyd and Nakahara 1 comment on and criticize a pair of
papers by Pareja 2' 3 in which a study of gas bubbles generated by annealing in silver
bicrystalline films is reported. The quantitative analysis of the gas bubbles showed a
dependence of their density on the annealing temperature and on the bicrystal
thickness, and a relationship between these parameters and the mean size of the gas
bubbles was also found. To explain the experimental results, a growth mechanism
for gas bubbles by the diffusion of gaseous impurity-vacancy pairs was proposed,
suggesting that gaseous impurities trapped in the film may be points for gas bubble
nucleation.
In the first place, we must point out that this growth mechanism is equivalent to
one proposed by Lloyd and Nakahara 4' 5 based on the annihilation of excess
vacancies at pre-existing voids by the diffusion of vacancies to these voids, since the
mechanisms are only different in the manner of the diffusion of vacancies to voids.
Both mechanisms result in the annihilation of vacancies at gas bubbles or voids.
Pareja never stated that the defects observed in his experiments were of the
same kind or the same origin as those observed by Lloyd and Nakahara 6 and
Andrew and Krasevec 7 because he had not observed pre-existing voids before
annealing. Nevertheless, the defects were certainly identified as real gas bubbles, i.e.
large voids containing a considerable quantity of gas from gaseous impurities
trapped in the bicrystalline film, on the basis of the diffraction contrast of the
electron microscopy image and of the stability at high temperatures which was
reported 2. It was evident that the large gas bubbles observed (about 13 nm or
greater) were in the bicrystal interface. Hence it might be accepted that they are
formed as a result of a simple coalescence of trapped gas at the welded interface,
especially water vapour from water adsorbed during the detachment of the film from
the substrate as carried out by Lloyd and Nakahara. However, we must point out
three objections to this argument. First, a large number of gas bubbles with much
smaller size, which may be as small as about 3-4 nm, was observed away from the
midplane of the bicrystals, i.e. the interface, as shown in ref. 2, Fig. 1. Secondly,
similar gas bubbles have been observed after welding of gold epitaxial film in air as
well as in situ under vacuum at pressures of the order of 10- 5 Pa s. Thirdly, the clear
correlation found between the gas bubble density and the bicrystal thickness as well
as between the mean size of the bubbles and the bicrystal thickness also confirm that
0040-6090/82/0000-0000/$02.75 © Elsevier Sequoia/Printed in The Netherlands