ISSN 10637397, Russian Microelectronics, 2015, Vol. 44, No. 2, pp. 132–138. © Pleiades Publishing, Ltd., 2015.
Original Russian Text © Yu.A. Novikov, 2015, published in Mikroelektronika, 2015, Vol. 44, No. 2, pp. 152–158.
Modern microelectronics and nanotechnology can
not progress without the measurement of linear dimen
sions. These technologies deal with elements whose
sizes vary from 1 nm to 100
m. Among devices used to
visualize objects having such dimensions, a scanning
electron microscope (SEM) stands out due to the fol
lowing reasons [1–7]. First, the SEM covers the desired
range of dimensions [6, 7]. Second, for the SEM oper
ating in the mode of collecting secondary slow electrons
(SSE mode), calibration test objects were fabricated
and calibration methods were developed [6–9]. Third,
methods for measuring the size of elements of micro
and nanostructures were elaborated to cover almost the
entire desired range of dimensions [6, 7, 10, 11].
Nevertheless, the SEM has some drawbacks, such
as vacuum measurement and the uncontrolled slope of
the sample under study. The use of a vacuum can be
both an advantage and a disadvantage, since a great
number of technological processes in micro and nan
otechnologies are conducted in a vacuum. However,
the uncontrolled slope of the sample is a significant
disadvantage. It is generally disregarded with the
explanation that the effect of the slope on the SEM
calibration and measurement of micro and nano
structure elements is low and does not exceed com
mon measurement errors. This, however, requires
proof. A method for experimental determination of
the test object slope is needed.
This paper presents such a method, along with the
results of investigating the effect of the test object’s
slope on the SEM calibration.
2. THEORY OF THE METHOD
Today, the SEM operating in the SSE mode is cali
brated using test objects in the form of relief pitch
structures [6–9, 12, 13]. Hills and valleys forming such
structures have a trapezoidal profile and high inclina
tion of the side walls [8, 9, 14]. Figure 1 shows the pro
file of the pitch structure for such a test object,
together with the model of the signal received by the
SEM operating in the SSE mode. The vertical dashed
lines show the connections between the boundary
points of the profile (see Fig. 1a) and the reference
points (see points
in Fig. 1b) of the signal.
If the conditions
is the (effective) diameter [15, 16] of an
electron probe and
is the inclination angle of the side
wall of a hill or valley with respect to the structure surface
normal (see Fig. 1a), then the parameters of the SSE sig
nal, the structure, and the SEM are related as follows:
is the pixelsize on the SSE image.
tan , , ,
LR pt pt
Effect of Test Object Slope on SEM Calibration
Yu. A. Novikov
A.M. Prokhorov General Physics Institute, Russian Academy of Sciences, Moscow, Russia
Received March 31, 2014
—The effect of the slope of a test object with a trapezoidal profile and high inclination of side walls
on the calibration of a scanning electron microscope (SEM) is investigated. Being used as an attested param
eter for the structure pitch or dimensions of upper and lower bases of hills and valleys, the slope of the test
object does not affect the calibration of the SEM. However, when using the projection of the inclined side wall
of hills and valleys onto the structure base to calibrate the SEM, the slope of the test object leads to a system
atic error in determining the increase (of the pixelsize), which is several times larger than the random error.
A method for determining the test object slope is developed that allows one to measure the slope, as well as
evaluate and eliminate the systematic inaccuracy.