106
4 Ultrathin Layers
chlorocomplexes was found to adsorb on the electrode that was reduced only later in
the period meant for the UPD layer formation, hence contributing to an unexpectedly
fast growth of the deposit [128].
The variation in surface roughness of SLRR layers with deposition cycle number is
somewhat similar to EC-ALD layers. When direct exchange takes place, the surface
roughness remains close to that of the substrate for a couple of SLRR layers. In
this case, the surface roughness increment is very low, and a nearly planar deposit
is obtained even if the same metal is deposited as three-dimensional nuclei during
d.c. deposition (like Pd with Cu sacrificial layer [143]). However, after a number
of cycles, roughening can take place, and the metal growth becomes faster than
what can be anticipated in the monolayer-wise growth process. The deposit is more
prone to develop a significant roughness as compared to the substrate if there is
a lattice mismatch between the deposit and the substrate. For instance, Au SLRR
on Au substrate causes much smaller surface roughening than the same process on
a Pt substrate [110]. The threshold thickness where roughening takes place varies
from one material to another. For H-adsorption-mediated Pd growth, the turnover was
found to be around 20 deposition cycles [113]. Concerning the SLRR growth of Pt, it
is the common experience for both classical [109, 112, 124, 128] and electroless [116]
processes that Pb as a sacrificial metal promotes a nearly even Pt layer growth, while
a Cu replacement leads to a progressive growth of the surface roughness practically
from the first cycle (see also Fig. 4.10).
Although it is not clearly evidenced, it is logical to assume that alloy layers
obtained with SLRR by using different deposited metals in the subsequent cycles
exhibit neither fully random nor completely ordered structure. In accord with the
original concept [12], such deposits may work as a bifunctional catalyst due to the
atomic arrangement at the surface that does not exist in conventional alloys. Mixed
Pt/Ru atomically layered catalysts proved to have beneficial properties for methanol
oxidation [12]. It is likely that for alloys produced with alternating replacement with
various growing metals at uneven cycle ratio, the minority component forms twodimensional island-like features in the matrix composed of the majority component
[107]. For a regular alteration of the components deposited from one growth layer to
another, an atomically layered structure was presented [129]. This picture shows the
most probable (dominant) type of atomic arrangement; nevertheless, the roughening
process is likely to induce irregularities in the deposit.
The inclusion of an intermediate layer between the substrate and the deposited
layer strongly impacts the deposit structure. Graphene-covered Au leads to a different
structure of Pt than the pure gold substrate [135], mostly due to the compressive strain
of the compact atomic spacing of the graphene sheet. Interestingly, the inclusion of
an intermediate graphene layer proved to be transparent in the sense that both UPD
on gold was possible to produce the intermediate layer and the Pt replacement could
be carried out in the usual way.
4 Ultrathin Layers
chlorocomplexes was found to adsorb on the electrode that was reduced only later in
the period meant for the UPD layer formation, hence contributing to an unexpectedly
fast growth of the deposit [128].
The variation in surface roughness of SLRR layers with deposition cycle number is
somewhat similar to EC-ALD layers. When direct exchange takes place, the surface
roughness remains close to that of the substrate for a couple of SLRR layers. In
this case, the surface roughness increment is very low, and a nearly planar deposit
is obtained even if the same metal is deposited as three-dimensional nuclei during
d.c. deposition (like Pd with Cu sacrificial layer [143]). However, after a number
of cycles, roughening can take place, and the metal growth becomes faster than
what can be anticipated in the monolayer-wise growth process. The deposit is more
prone to develop a significant roughness as compared to the substrate if there is
a lattice mismatch between the deposit and the substrate. For instance, Au SLRR
on Au substrate causes much smaller surface roughening than the same process on
a Pt substrate [110]. The threshold thickness where roughening takes place varies
from one material to another. For H-adsorption-mediated Pd growth, the turnover was
found to be around 20 deposition cycles [113]. Concerning the SLRR growth of Pt, it
is the common experience for both classical [109, 112, 124, 128] and electroless [116]
processes that Pb as a sacrificial metal promotes a nearly even Pt layer growth, while
a Cu replacement leads to a progressive growth of the surface roughness practically
from the first cycle (see also Fig. 4.10).
Although it is not clearly evidenced, it is logical to assume that alloy layers
obtained with SLRR by using different deposited metals in the subsequent cycles
exhibit neither fully random nor completely ordered structure. In accord with the
original concept [12], such deposits may work as a bifunctional catalyst due to the
atomic arrangement at the surface that does not exist in conventional alloys. Mixed
Pt/Ru atomically layered catalysts proved to have beneficial properties for methanol
oxidation [12]. It is likely that for alloys produced with alternating replacement with
various growing metals at uneven cycle ratio, the minority component forms twodimensional island-like features in the matrix composed of the majority component
[107]. For a regular alteration of the components deposited from one growth layer to
another, an atomically layered structure was presented [129]. This picture shows the
most probable (dominant) type of atomic arrangement; nevertheless, the roughening
process is likely to induce irregularities in the deposit.
The inclusion of an intermediate layer between the substrate and the deposited
layer strongly impacts the deposit structure. Graphene-covered Au leads to a different
structure of Pt than the pure gold substrate [135], mostly due to the compressive strain
of the compact atomic spacing of the graphene sheet. Interestingly, the inclusion of
an intermediate graphene layer proved to be transparent in the sense that both UPD
on gold was possible to produce the intermediate layer and the Pt replacement could
be carried out in the usual way.