en: {nie the case of Me-DLC an especially adapted method has been used, which makes use of the fact that
Pole particles are more or less of spherical shape and do not penetrate each other due to the high surface
US Be yg energy of the metallic phase. The method has been described in detail in (7) and for different STM
is ei measurements individual tip radii between 6 +2 and 131 +7 A have been found. Using these tip ra-
heel br, dii particle radii could be determined (see figure 2) which correspond well to results of other ana-
St fae lytical methods. A careful tip radius determination was found to be very important since tip radii
sole Lina and particle radii are of the same order of magnitude, leading to severe errors if not considered.
wa Ay The particle distance determination, too, is affected by the finite size of the tip because small parti-
nan cles or particles lying deep between other ones are partly or completely hidden due to the tip-
+ : indispen- surface convolution. Hence nearest neighbour distances are replaced by the next to nearest neigh-
“convoution bour distances, leading to a shift of the distance distribution to higher values. Even if the tip radius
. of curvature is known and image deconvolution can be performed, reconstruction of hidden parti-
ft cles and their corresponding distances is impossible since the STM image does not contain infor-
mation about them.
4 Nevertheless, to get useful distance values out of
—_ the STM images, simulations of the tip surface
e150| 7 1SAXS convolution by Monte-Carlo methods can be per-
2ical protection ne i 15M corr formed to compute the fraction of visible and in-
+ ve DC- c visible particles and their respective distances as a
n atmosphere, : 90 + function of the tip radius of curvature, the particle
illic or carbidic 60 radius distribution and the apparent particle dis-
rts DLC ma- a30 tance-distribution of the sample. As a result of a
«30 he varied Po. self-consistent, iterative procedure a correction
ers in a Wide 0 10 20 30 40 50 factor can be calculated and applied to the experi-
nge of particle Au [at%] mental mean distance values, taking the conceal-
a tum strongly . N } i ment of particles into consideration. It was found
ectancal and re BT and Mon Aud LC Coatings that, depending on tip radii and metal content of the
’herefore, par- on, compared by results of SAXS and TEM. samples, apparent particle distances are about 15%
ıterest for the to 35% too high due to hidden particles, and a dis-
tinct improvement can be achieved using the cor-
in air using rected values. The results are shown in fig 3. The remaining differences between STM results and
Loe the metal- reference methods may result from the assumptions of the Monte-Carlo model but also real devia-
‚m convolution tions between distances at the surface (STM) and within the bulk material (TEM, SAXS) may exist.
N to increased A detailed description of the Monte-Carlo model and the respective results can be found in (7). The
+ oer nd complete results for Au-, Pt-, W- and Fe-DLC have be described elsewhere (8).
re radius of
jetermine indi-
now de tip 12 Material Contrast by Barrier Height Imaging
nn dpe by
v . on ole Imaging of the tunnelling barrier height (BH) in STM is one method which may provide contrasts
Cg shape between different materials. Using Tersoff and Hamann’s approach for the tunnelling current (9)
| ons of the one obtains the following derivation of the tunnel current I with respect to the tunnel distance z:
elie (dUdz) ~ 1 sqrt(¢) cos(0r)
i from the where ¢ = _ (®sp + Osampıe) = the effective tunnelling barrier height and a is the local tilt angle of the
ne yes, always surface (10,11). Modulation of the tunnel distance with subangstrom amplitude and some kilohertz
ang frequency results in a corresponding modulation of the tunnel current, whose amplitude can be de-
me tected by lock-in technique. The lock-in signal is proportional to the square root of the local BH
er umatitgl yielding a material specific contrast. Due to the cos(o)-term it is mixed-up with a topographic con-
; aes In trast. which lower the apparent BH at strongly tilted surfaces. An example can be seen in fig. 4.
5
