Vacuum-ultraviolet spectroscopy and quantum cutting for Gd
3
in LiYF
4
R. T. Wegh,* H. Donker, and A. Meijerink
Debye Institute, Utrecht University, P.O. Box 80 000, 3508 TA Utrecht, The Netherlands
R. J. Lamminma
¨
ki and J. Ho
¨
lsa
¨
Department of Chemistry, University of Turku, FIN-20014 Turku, Finland
Received 27 March 1997
A systematic spectroscopic study of the 4 f
7
energy levels of Gd
3+
in LiYF
4
in the vacuum-ultraviolet
spectral region (50 000 – 70 000 cm
-1
) is reported. Using energy-level calculations, all observed spectral lines
could be assigned to free-ion term symbols including term symbols with unusually high L and J , e.g., a
2
Q
23/2
level around 67 000 cm
-1
. From the
6
G
J
levels around 50 000 cm
-1
quantum cutting or two-photon lumi-
nescence, photon-cascade emission is observed: the emission of a red photon due to the
6
G
J
→
6
P
J
transition
is followed by the emission of an ultraviolet photon due to the
6
P
J
→
8
S
7/2
transition.
S0163-18299701646-9
I. INTRODUCTION
In the past decades experimental and theoretical work by
scientists like Dieke,
1
Carnall et al.,
2
and many others has
provided an overview of the 4 f
n
energy levels of all rare-
earth ions in the infrared, visible, and ultraviolet spectral
region. Energy-level calculations up to 50 000 cm
-1
have
been performed for all rare-earth ions and most of the energy
levels have been observed experimentally.
In spite of the large number of papers on 4 f
n
energy
levels of rare-earth ions, reports on 4 f
n
energy levels in the
vacuum-ultraviolet spectral region VUV; E 50 000 cm
-1
,
200 nm are scarce. For the strong parity allowed
4 f
n
→4 f
n -1
5 d transitions on the other hand, the spectra in
the VUV region have been studied for most rare-earth ions in
LaF
3
, CaF
2
, and LiYF
4
, starting with the pioneering work by
the group of Yen.
3,4
Only for a few rare-earth elements some
4 f
n
levels in the VUV have been calculated and/or
observed.
5–9
One can think of two reasons for this lack of
effort in the VUV region on 4 f
n
→4 f
n
transitions: 1 It is
difficult. Measurements on the weak parity forbidden intra-
configurational 4 f
n
→4 f
n
transitions require special setups
for luminescence spectroscopy in the VUV and calculations
including levels above 50 000 cm
-1
require larger matrices.
Furthermore, the measurements on the 4 f
n
→4 f
n
transitions
can be complicated by the difficulty of discriminating them
from parity-allowed 4 f
n
→4 f
n -1
n ' l background bands.
Two-photon absorption spectroscopy has been used success-
fully to overcome this problem,
10
but to date this has not
been extended into the VUV. 2 There are no applications
that require knowledge on the 4 f
n
levels of rare-earth ions in
the VUV.
The latter point has changed recently. The development of
phosphors for excitation in the VUV has become an impor-
tant new challenge in the field of luminescent materials re-
search. VUV phosphors are required for application in
mercury-free fluorescent tubes and in plasma display panels.
In these devices a noble-gas discharge generates VUV radia-
tion. The xenon dimer discharge, which yields a broad band
in the VUV with the maximum at 172 nm xenon dimer,
gives the highest efficiency of all noble gases, but it is still
less efficient than the conventional mercury discharge. The
phosphors used in mercury discharge fluorescent tubes have
quantum efficiencies close to 100%. Therefore, to make a
mercury-free fluorescent tube competitive, a phosphor with a
quantum efficiency higher than 100% is required. In other
words, more than one visible photon should be obtained per
absorbed VUV photon. One of the challenges is to find such
a VUV phosphor, a so-called quantum cutter. The high en-
ergy of the VUV photons from a xenon dimer discharge
makes it possible in theory to obtain quantum cutting or
two-photon luminescence, photon-cascade emission in the
visible. For rare-earth ions the phenomenon of quantum cut-
ting is known,
11,12
but based on the current knowledge no
efficient quantum cutter in the visible is possible.
13
Our research program on finding an efficient quantum cut-
ter involves three stages. First, the energy levels in the VUV
region will be resolved for a number of rare-earth ions. Next,
it will be investigated if efficient visible quantum cutting
from one of the VUV levels is possible. Finally, promising
ions will be incorporated in various host lattices to find a
stable VUV phosphor with a high close to 200% quantum
efficiency in the visible. The first ions to be investigated will
be those ions that are able to give an efficient emission in the
visible, either directly or after energy transfer, e.g., Sm
3+
,
Eu
3+
, Gd
3+
, Ho
3+
, and Er
3+
. To resolve the energy levels in
the VUV region the lanthanides are incorporated in fluoride
lattices e.g., LiYF
4
and LaF
3
. In these host lattices the op-
posite parity states 4 f
n -1
5 d and charge transfer that can
interfere with the energy levels of the 4 f
n
configuration are
at the highest possible energies.
To investigate if efficient quantum cutting in the visible
can be achieved, the emission spectra under VUV excitation
will be analyzed. For efficient quantum cutting in the visible
two conditions must be fulfilled. First, to get emission from a
high-lying energy level, the gap to the next lower level
should be large enough to prevent multiphonon relaxation.
14
Second, the branching ratio i.e., the way the total emission
intensity is divided over the various transitions in the IR,
visible, and UV spectral region must be appropriate. This
PHYSICAL REVIEW B 1 DECEMBER 1997-I VOLUME 56, NUMBER 21
56 0163-1829/97/5621/138418/$10.00 13 841 © 1997 The American Physical Society