Luminescence of mercuric iodide crystals
I. Kh. Akopyan, B. V. Bondarenko, O. N. Volkova, B. V. Novikov, and T. A. Pavlova
Physics Research Institute, St. Petersburg State University, 198904 Petrodvorets, Russia
Submitted August 2, 1996
Fiz. Tverd. Tela St. Petersburg 39, 67–73 January 1997
Low-temperature 4.2–130 K photoluminescence spectra of HgI
2
crystals have been measured
in the 540–700 nm region. An analysis of the characteristics intensity vs temperature
and excitation power relations, afterglow times, excitation spectra of the 560, 620, and 635 nm
emission bands suggests the following assignments: the 560 nm band is due to radiative
annihilation of excitons bound to mercury vacancies, and the ‘‘red’’ emission originates from
recombination of free 620 nm and donor-localized 635 nm electrons with a hole-
filled acceptor level. The energies of the corresponding donor and acceptor levels have been
estimated. New emission bands at 540, 545, and 575 nm have been discovered, and their origin
discussed. © 1997 American Institute of Physics. S1063-78349701201-X
Despite many years of studies of the optical spectra of
tetragonal -HgI
2
Refs. 1–8, the broad luminescence
bands observed in them have not yet received unambiguous
assignment, and the nature of the defects involved remains
unclear. This problem gains, however, in importance in con-
nection with the application potential of HgI
2
which is pres-
ently opening up. Mercuric iodide is the best material for
nuclear radiation detectors that do not need cooling, however
their quality has not reached a high enough level because of
the low carrier mobility, which is due to high defect concen-
trations in samples. Probing the nature of these defects by
photoluminescence techniques has become a major focus of
activities in recent studies of this compound.
7–11
The low-temperature, luminescence spectrum near the
fundamental absorption band edge of red -HgI
2
consists of
strong narrow-band emission in the 529–540 nm-region and
of broad bands peaking at 560 and 630 nm Refs. 1 and 2.
We have made recently a comprehensive study of the 529–
540 nm luminescence, which is due to radiative annihilation
of free excitons at 77 K and of excitons bound to intrinsic
point defects at 4.2 K, and to their interaction with lattice
vibrations. This work deals with a study of the 560 and 630
nm broad bands and discusses the nature of the defects re-
sponsible for their presence in the luminescence spectra of
HgI
2
. We have also observed and studied luminescence
bands of mercuric iodide near 510, 540, and 575 nm.
We investigated single crystals produced a by isother-
mal ( T =25 °C evaporation of an acetone solution of HgI
2
Ref. 12, b by vapor-phase transport,
13
and c by the tem-
perature oscillation technique from the vapor phase.
14
We
studied also the luminescence spectra of HgI
2
crystals con-
tained in multicomponent systems which form in solid-phase
chemical reactions between HgI
2
and AgI crystals.
15
The
spectra were excited by a mercury lamp, a Cd laser
(
exc
=441.6 nm and a N
2
laser (
exc
=337.1 nm.
1. THE 560 NM-BAND
This band with a halfwidth of about 60 meV at
T =4.2 K is observed in luminescence spectra of practically
all crystals of different origin, and is the strongest in crystals
grown by vapor-phase transport with supersaturation in
iodine
7
. As the crystal temperature is increased, the band
gradually weakens in intensity but remains visible in the
spectra up to 110 K Fig. 1. While varying insignificantly
below 30 K, above it the intensity exhibits an exponential
behavior as a function of inverse temperature, with
E =60 meV inset to Fig. 1. Within this interval, the ther-
mal quenching of the 560 nm-band is stronger than that of
the A band of free-excitons.
The afterglow times were shown not to exceed 60 ns at
T =4.2 K. Measurements made with a delay time of 50 ns
did not reveal any change in band shape or position com-
pared to the stationary spectrum.
Increasing the excitation power at T =4.2 K resulted in a
noticeable shift of the band maximum toward higher ener-
gies, from 560 nm under mercury lamp pumping to 556 nm
when a focused N
2
laser was used, i.e., by 20 meV.
Our absorption measurements showed a structure reso-
nant with the 560 nm band. Studies of absorption in this
spectral region present experimental difficulties because of
the absence of sufficiently thick uniform samples. While
crystals grown by the same technique may already differ in
stoichiometry, the stoichiometry of the surface may be quite
different from that in the bulk. Besides, the conditions of
storage may affect the stoichiometry and, hence, the concen-
tration and character of the intrinsic defects. Therefore, in
some cases we had to obtain absorption spectra on samples
made up of an array of several freshly cleaved specimens,
which had been first selected according to their luminescence
spectra. Curve 1 in Fig. 2 demonstrates the absorption spec-
trum of a sample exhibiting a strong 560 nm-luminescence
band curve 1a throughout its thickness. Crystals with a
weak 560 nm luminescence band curve 2a or no band at all
do not have an absorption structure at 556 nm curve 2.
Taking into account the short lifetime and the existence
of resonant absorption, we believe the radiative decay of
excitons bound to deep centers to be the most probable
mechanism responsible for the emission band near 560 nm.
This assignment is supported by the excitation spectra;
16
in-
deed, in contrast to emission in the red region which is ex-
cited only by interband transitions, the 560 nm band is ob-
58 58 Phys. Solid State 39 (1), January 1997 1063-7834/97/010058-06$10.00 © 1997 American Institute of Physics