Page 133 - Semiconductor Manufacturing Handbook

Geng(SMH)_CH10.qxd 04/04/2005 19:46 Page 10.6

ION IMPLANTATION AND RAPID THERMAL PROCESSING

10.6 WAFER PROCESSING

along the width of the ribbon. While this architecture has gained some acceptance, it is generally
accepted that fundamental limitations to low-energy beam transport exist in this configuration that
make the low-to-sub-keV energy ranges impractical for production operation.
In addition to the drift mode of operation that is most common for high-current tools in the
energy ranges of production interest today, most high-current tools can also be operated in a decel-
eration mode (sometimes also referred to as a decel or differential mode) wherein the ion beam is
transported through the majority of the beamline at an energy much higher than the final desired
energy, and then decelerated by means of an electrostatic field to the final energy just before the
beam is incident on the wafer. 9,10 The deceleration mode is typically used only when the ion beam
transport through the beamline is at an energy that is so low as to make beam current losses due to
space charge beam expansion very problematic from a productivity standpoint. By allowing the
beam to propagate through the beamline at a higher energy, the space charge beam expansion and
losses are mitigated. The decel mode of operation can be employed on both fixed-spot and fixed-
ribbon beam beamlines. The deceleration typically occurs in a single stage at or just following the
resolving aperture on the fixed-spot beam tools. On fixed-ribbon beam tools, the deceleration can
occur in a single stage in the vicinity of the resolving aperture, or in multiple stages, the first of
which is in the vicinity of the resolving aperture and the second following the parallelizing magnet,
immediately before the wafer.
Most concerns related to the deceleration mode have to do with the possibility of some beam flux
reaching the wafer at energies other than the intended final energy. This so-called energy contami-
nation is usually a result of charge exchange reactions that take place between fast beam ions and
slow background gas neutral atoms or molecules. The charge exchange reaction results in a fast neu-
tral beam atom and a slow background gas ion. The fast neutral beam can no longer be influenced
by the decelerating electrostatic field that is necessary in the deceleration mode. Because of this, any
fast neutral beam created before deceleration and within the line of sight of the wafer will arrive at
11
the wafer with an energy that is higher than the intended energy. The typical amount of energy con-
tamination in high-current tools making use of the deceleration mode in production today is of the
order of 1 percent of the intended total dopant dose. Depending on the sensitivity of the processing
step to this amount of energy contamination, or to day-to-day variations in this amount of energy
contamination, this may be a serious problem.

10.2.4 High-Energy Beamlines
The primary objective of high-energy implanters is to deliver up to hundreds of microampere beams
at energies up to several MeV. There are two fundamental beamline approaches to achieving this
objective. Most commercial high-energy systems today make use of a radio frequency (RF) linear
12
accelerator (referred to as a linac) to deliver the MeV range of energies. The commercial RF linac
relies on a conventional ion source and analyzer magnet, not unlike those found in a high-current
beamline, to produce a dc beam of up to several milliamperes at approximately 90 keV. This dc beam
is then injected into a series of 8 to 12 electrodes connected to RF resonators operating at 13.56
MHz that further accelerate the beam. The resonators serve to provide accelerating voltages of up to
approximately 80 kV on each electrode. The first two RF electrodes bunch the beam into discrete
packets and provide some acceleration, thereby forming a pulsed ion beam. Each beam packet is then
further accelerated by the subsequent RF electrodes, each of which provides the opportunity to accel-
erate the beam both as it enters and leaves the electrode (by virtue of the fact that the electric field
changes polarity at the RF while the beam packet is transiting the length of the electrode—a so-called
push-pull technique). Each electrode in the series has a length that is chosen to provide an optimum
transit time for each beam packet (which moves faster after each subsequent acceleration stage). A pair
of electrostatic quadrupole lenses is placed between each RF stage to provide additional control
over the beam size and shape. Following the acceleration to the final energy in the linac stage of
the beamline, the beam passes through an electromagnet (typically referred to as the final energy
magnet, or FEM) that deflects, disperses, and focuses the beam through an angle of approximately 40°
(again according to Eq. (10.4)) and ensures that only ions of the desired momentum are allowed to pass

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)
Copyright © 2004 The McGraw-Hill Companies. All rights reserved.
Any use is subject to the Terms of Use as given at the website.