Solid-State Nanopores: Methods of Fabrication and Integration, and Feasibility Issues in DNA Sequencing - Nanopores: Sensing and Fundamental Biological Interactions - page 182

This equation can be used to interpret all behaviors described earlier (Storm
et al. [ 6 ], Wu et al. [ 8 ]). Since r << r d during the entire shrinking process, Eq. (8.3)
can be reduced to

r ¼g = 4 : (8.4)

According to Eq. (8.4), the radius (or diameter) of a pore in a fluidized membrane
will decrease linearly with time if the starting diameter is less than the membrane
thickness. The rate at which it shrinks is determined by the ratio of surface tension
to viscosity, both of which are dependent on the material and temperature (and
other physical conditions). However, the dynamics is universal.

8.2.2 TEM “Drilling”

It started out as a curiosity while considering the simplified surface-tension model
in Fig. 8.2 . According to Fig. 8.2b , the energy barrier to thermally nucleating a pore
in an uniform membrane of thickness h is ~ h 3 . Thus in very thin membranes, it
should be possible to nucleate a pore by locally heating up the membrane. This
hypothesis was quickly put to the test, and it was quickly found (Storm et al. [ 6 ])
that indeed it is possible to form a nanopore in a thin uniform SiO 2 membrane.
The exact mechanism of the pore forming under a local focused e-beam is still
uncertain. The careful element analyses (Wu et al. [ 11 ] and Kim et al. [ 12 ]) of
the material near the pore showed the loss of oxygen. Thus it is common in the
solid-state nanopore field to consider the pore forming mechanism under a focused
e-beam to be equivalent to “drilling”, or removal of materials locally (Fig. 8.4 ).

Fig. 8.4 TEM images of two nanopores formed by TEM “drilling.” The membrane is Si 3 N 4 , with
thickness ~20 nm. The TEM used is a JEOL 2010 F system with a field-emission source operating
at 200 keV. The scale bars ¼ 5 nm. (Images taken by V. Balagurusamy, unpublished, 2010)