1063-7397/05/3405- © 2005 MAIK “Nauka /Interperiodica”
Russian Microelectronics, Vol. 34, No. 5, 2005, pp. 316–324. Translated from Mikroelektronika, Vol. 34, No. 5, 2005, pp. 375–384.
Original Russian Text Copyright © 2005 by Borodovskii, Buldygin, Tokarev, Chernyavskii.
Microwave measurements of excess-carrier life-
time were ﬁrst made on germanium in 1959 by
. The property was evaluated from
the time variation of transmitted microwave power,
, under pulsed injection, with the specimen ori-
ented perpendicular to the waveguide axis. In 1962,
Deb and Nag  proposed a reﬁned version of the
method in which carrier lifetime is measured by
reﬂected rather than transmitted power. It has gained
wide acceptance as a tool for characterization of sil-
icon wafers in semiconductor manufacture. Relevant
standards have been deﬁned in the United States
(ASTMF 1535-94) and Japan (JEDA-53).
On the other hand, early theoretical papers on the
method, such as the ones by Atwater , expressed
some doubt about its reliability on the grounds that it
was difﬁcult to identify the conditions under which
has a linear relationship to the excess-conductiv-
; otherwise, the calculation of
was a severe problem. Jacobs
 carried out
extensive calculations and measurements for the case
where a specimen completely blocked a waveguide.
Conditions for reliable determination of carrier lifetime
were thus ascertained from data on germanium and sil-
icon specimens differing in thickness and resistivity.
Moreover, close agreement was achieved between
direct-current and microwave measurements on a spec-
imen. Bhar  addressed the microwave measurement
of diffusion length and lifetime in a transversely ori-
ented specimen under pulsed illumination of its end.
The reﬂection version of the method as applied to sili-
con wafers was subsequently investigated in many the-
oretical and experimental studies. A large number of
relevant references are given by Kunst and Beck  and
Schöfthaler and Brendel .
The case of a specimen located outside a resonator
was considered by V.V. Akhmanov
. An advan-
tage of this arrangement is that it can be used with
ingots as well as specially prepared specimens.
This paper is a continuation of the work by Boro-
, in which a new method was pro-
posed for the microwave measurement of bulk life-
time in monocrystalline silicon ingots. It is again
assumed that the excess conductivity as measured by
absorbed microwave power is proportional to the
total number of excess carriers excited by a light
pulse. An output pulse from a detector that senses the
transmitted microwave power is determined by the
sum of changes in reﬂected and absorbed microwave
power. As we shall show, the reﬂection mode of mea-
surement does not always ensure that the magnitude
of a photoconductivity response is proportional to
the total number of excess carriers in the volume
being examined, due to front-surface reﬂection. This
effect is considered in the present study. We shall
also report on predictions and measurements con-
cerning the decay of a photoconductivity response
for different light-pulse widths. The aim is to iden-
tify the conditions for reliable measurement of bulk
carrier lifetime in undoped and lightly doped ﬂoat-
zone silicon ingots. The decay of excess-carrier den-
sity is calculated by approximate formulae derived
for a semi-inﬁnite specimen and linear recombina-
tion in the volume and at the surface [10, 11].
Method for the Microwave Measurement of Carrier Lifetime
in Lightly Doped Silicon Ingots
P. A. Borodovskii, A. F. Buldygin, A. S. Tokarev, and E. V. Chernyavskii
Institute of Semiconductor Physics, Siberian Division,
Russian Academy of Sciences, Novosibirsk, Russia
Received December 24, 2004
—A further investigation is conducted into a new microwave method for evaluating bulk lifetime in
silicon ingots . It is shown that for lightly doped ingots transmission and reﬂection microwave measurements
yield different values of the decay time of photogenerated excess-carrier density. The discrepancy is attributed
to the inﬂuence of surface recombination. Measurements are compared with theoretical predictions for a semi-
inﬁnite semiconductor specimen. Data are presented on bulk lifetimes in ten high-quality silicon ingots grown
by the ﬂoat-zone process and doped with phosphorus by neutron transmutation doping. The lifetimes are mea-
sured by microwave transmission or reﬂection and by probe injection.
MATERIALS AND MICROSTRUCTURE