12.2 Two-Plate Thin-Layer Cells for Electron Microscopic …
437
12.2.3 In Situ Electroplating Studies with TEM
The first study dealing with the in situ TEM observation of the continuous deposit
formation with the “current in solution layer plane” configuration referred to Ni
deposition from NiCl 2 solution [36]. The voltage was controlled in a two-electrode
system during deposition and subsequent dissolution of Ni. The experimental finding
was difficult to elucidate because of the artefact that the silicon nitride window did
not prove to be a good enough insulator. This led to the unexpected finding that
Ni nucleation could take place in advance of the growth front of the continuous
deposit being present on the cathode itself. Hence, the “electrode” and “substrate”
was not essentially the same. The propagation of the growth front during the noninhibited growth of the deposit showed a large lateral inhomogeneity, consistently
with the dendritic morphology obtained. Upon dissolution, the elctropolishing conditions could be observed in the sense that the dissolution front propagation was much
more uniform than the growth front.
The in situ TEM observation of Pb growth and dissolution was partly contradictory
to the Ni deposition/dissolution study [42]. Here, the goal was the study of dendrite
formation. What was found is that the dendrite growth and dissolution were asymmetrical in the sense of the thickness and length changes. Upon growth, the dendrites grew
preferentially in the current direction with nearly proportional thickening. However,
the dissolution took place through a remarkable thinning before being shrunk along
the perpendicular-to-substrate direction. This observation is highly in contrast to
electropolishing theories and can probably be explained with the solution concentration changes during the process. Since the solution used was saturated with respect
to lead nitrate, the partial depletion between the dendrites might have been provided
more favourable conditions for dissolution than the end of the dendrite tip where the
saturation concentration could be spontaneously restored much more easily. Similar
results were obtained in another study [33] where the reactant concentration was
far from the saturation level and surfactant was also applied to regulate the deposition. The detailed analysis of the branching dendrite showed that the deposition
and dissolution process of the side branches are nearly symmetrical concerning the
length/width ratio, but the main dendrite anchoring directly on the substrate tends to
be thinned before being shortened. It was also shown that the decrease in the reactant
concentration promotes both the dendrite formation and the branching.
In studies dealing with Cu deposition, the general trends established for surface
roughness evolution were confirmed also at the nanoscale by imaging about 2-µmwide sections of the growing deposits [43, 44]. Deposits obtained with a relatively
small d.c. current were less rough that obtained with a rate close to the diffusionlimited growth. Pulse plating was shown to suppress the development of the roughness even at a high rate, and PbSO 4 applied as an inhibitor was shown to almost
completely eliminate the increase of surface roughness.
437
12.2.3 In Situ Electroplating Studies with TEM
The first study dealing with the in situ TEM observation of the continuous deposit
formation with the “current in solution layer plane” configuration referred to Ni
deposition from NiCl 2 solution [36]. The voltage was controlled in a two-electrode
system during deposition and subsequent dissolution of Ni. The experimental finding
was difficult to elucidate because of the artefact that the silicon nitride window did
not prove to be a good enough insulator. This led to the unexpected finding that
Ni nucleation could take place in advance of the growth front of the continuous
deposit being present on the cathode itself. Hence, the “electrode” and “substrate”
was not essentially the same. The propagation of the growth front during the noninhibited growth of the deposit showed a large lateral inhomogeneity, consistently
with the dendritic morphology obtained. Upon dissolution, the elctropolishing conditions could be observed in the sense that the dissolution front propagation was much
more uniform than the growth front.
The in situ TEM observation of Pb growth and dissolution was partly contradictory
to the Ni deposition/dissolution study [42]. Here, the goal was the study of dendrite
formation. What was found is that the dendrite growth and dissolution were asymmetrical in the sense of the thickness and length changes. Upon growth, the dendrites grew
preferentially in the current direction with nearly proportional thickening. However,
the dissolution took place through a remarkable thinning before being shrunk along
the perpendicular-to-substrate direction. This observation is highly in contrast to
electropolishing theories and can probably be explained with the solution concentration changes during the process. Since the solution used was saturated with respect
to lead nitrate, the partial depletion between the dendrites might have been provided
more favourable conditions for dissolution than the end of the dendrite tip where the
saturation concentration could be spontaneously restored much more easily. Similar
results were obtained in another study [33] where the reactant concentration was
far from the saturation level and surfactant was also applied to regulate the deposition. The detailed analysis of the branching dendrite showed that the deposition
and dissolution process of the side branches are nearly symmetrical concerning the
length/width ratio, but the main dendrite anchoring directly on the substrate tends to
be thinned before being shortened. It was also shown that the decrease in the reactant
concentration promotes both the dendrite formation and the branching.
In studies dealing with Cu deposition, the general trends established for surface
roughness evolution were confirmed also at the nanoscale by imaging about 2-µmwide sections of the growing deposits [43, 44]. Deposits obtained with a relatively
small d.c. current were less rough that obtained with a rate close to the diffusionlimited growth. Pulse plating was shown to suppress the development of the roughness even at a high rate, and PbSO 4 applied as an inhibitor was shown to almost
completely eliminate the increase of surface roughness.