Browse Prior Art Database

Spot Encoding And Increased Speed of Electron-Beam Inspection of Semiconductor Wafers

IP.com Disclosure Number: IPCOM000120605D
Original Publication Date: 1991-May-01
Included in the Prior Art Database: 2005-Apr-02
Document File: 6 page(s) / 223K

Publishing Venue

IBM

Related People

Batchelder, JS: AUTHOR [+4]

Abstract

A technique is described whereby the speed of electron-beam (E-beam) inspection processing is increased so as to improve the rate of monitoring and inspection of semiconductor wafers for contamination. Also described is the use of a multiple scanned spot system to allow for simultaneous detection of multiple pixels.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 40% of the total text.

Spot Encoding And Increased Speed of Electron-Beam Inspection of
Semiconductor Wafers

      A technique is described whereby the speed of
electron-beam (E-beam) inspection processing is increased so as to
improve the rate of monitoring and inspection of semiconductor wafers
for contamination. Also described is the use of a multiple scanned
spot system to allow for simultaneous detection of multiple pixels.

      Although standard semiconductor inspection techniques have good
sensitivity in detecting contamination particles on monitor wafers,
the throughput can be improved.  The concept described herein
incorporates the use of a multiple spot E-beam to increase the
throughput.

      The multiple spot E-beam operates such that if a spot scanned
by the E-beam is more complicated than a uniform or gaussian cross
section, then it is possible to use a larger scanned spot without
losing signal to noise.  Multiple gun E-beams are not new; an example
is the high resolution cathode ray tube (CRT) displays for X-ray
imaging using eight independent E-beam guns.  Use can be made of
interfering beams, or by passing a beam through a specially designed
aperture.

      Fig. 1 shows a large scanning spot scanning across a small
particle. The resulting secondary electron signal versus time is
shown in Fig. 2. The limit to detectability is set by the statistical
noise of the secondary electron signal.  Fig. 3 shows a large spot
formed from a series of parallel lines.  The internal structure of
the spot is chosen to increase the amount of information returned by
the secondary electrons produced by the particle, as opposed to those
produced by the flat substrate.  A chirped grating is appropriate, so
that the secondary electron signal produced by the scanned spot, as
shown in Fig. 4, is at a DC level with a superimposed frequency
chirped signal.

      Fig. 5 shows the use of the E-beam gun with its beam focused
through the focus coils and the chirped grating, through the focus
and scanning secondary detector to the wafer.  The resulting signal
processing is similar to that done in radar to improve the signal to
noise of reflected beams off remote objects produced by a relatively
weak source.  A high-pass filter is first applied to the signal from
the detector which cuts off at a frequency that corresponds to the
dwell time of the scanned spot.  The signal then goes into a matched
filter, which compresses the component of the observed signal with
the frequency chirp associated with the particle signal.  The
increase in signal to noise using this technique is given by the
product of the duration of the chirp times the bandwidth of the
chirp.

      One example would be a five micro E-beam spot that passes
through a chirped grating in the beam column that has thirty grating
periods in it and whose periods of grating lines varies by a factor
of two across the grating.  The projected spot has a dwell time of
one microsecond.  Th...