Frequency upconversion in Er3+-doped fluoroindate glasses pumped at 1.48 μm

Article abstract of DOI:10.1103/PhysRevB.55.6335

We report on efficient frequency upconversion in Er³⁺-doped fluoroindate glass. The process is observed under 1.48 μm laser diode excitation and results in fluorescence generation in the range from ultraviolet to near-infrared radiation. The study was performed for samples containing 1, 2, and 3 ErF₃ mol % in the range of temperatures from 24 to 448 K. The upconverted signals were studied as a function of the laser intensity, and their dynamical behavior is described using a rate equation model which allows us to obtain the energy transfer rates between Er³⁺ ions in pairs and triads.

Full text of DOI:10.1103/PhysRevB.55.6335

PHYSICAL REVIEW B
VOLUME 55, NUMBER 10
1 MARCH 1997-II
Frequency upconversion in Er3؉-doped fluoroindate glasses pumped at 1.48 ␮m
´
G. S. Maciel and Cid B. de Araujo
´
Departamento de Fısica, Universidade Federal de Pernambuco, 50670-901 Recife, PE, Brazil
Y. Messaddeq
´
˜
Departamento de Quımica, Universidade do Estado de Sao Paulo, 14800-900 Araraquara, SP, Brazil
M. A. Aegerter
¨
¨
Institut fur Neue Materialien, Im Stadtwald, Gebaude 43, D66-123 Saarbrucken, Germany
͑Received 16 July 1996; revised manuscript received 15 October 1996͒
We report on efficient frequency upconversion in Er3ϩ-doped fluoroindate glass. The process is observed
under 1.48 ␮m laser diode excitation and results in fluorescence generation in the range from ultraviolet to
near-infrared radiation. The study was performed for samples containing 1, 2, and 3 ErF3 mol % in the range
of temperatures from 24 to 448 K. The upconverted signals were studied as a function of the laser intensity,
and their dynamical behavior is described using a rate equation model which allows us to obtain the energy
transfer rates between Er3ϩ ions in pairs and triads. ͓S0163-1829͑97͒07509-7͔
I. INTRODUCTION
In the present work we report results of our investigations
on the upconversion properties of Er3ϩ-doped fluoroindate
glass using the infrared radiation from a cw diode laser as the
excitation source. This work extends our previous room-
temperature studies7 for the whole range of temperatures
from 24 to 448 K.
In the past years fluoride glasses doped with rare-earth
͑RE͒ ions have received great attention due to the possibili-
ties of using these materials in numerous applications such as
the operation of upconversion lasers, superfluorescent
sources, and optical amplifiers, among others.1 Fluoride
glasses are particularly attractive hosts because they can be
fibered, maintaining a high intensity of pumping light over a
long interaction length, and large RE concentrations can be
incorporated easily into the host matrix. Furthermore, one of
the advantages of fluoride hosts is the low energy of its more
energetic phonons, which reduces the probability of mul-
tiphonon relaxation processes between the RE electronic
levels.
II. EXPERIMENTAL DETAILS
The glasses studied have the following % molar
composition: ͑39Ϫx͒InF3-20ZnF2-16BaF2-20SrF2-2GdF3-
2NaF-1GaF3ϪxErF3 ͑xϭ1,2,3͒. The samples preparation
procedure is briefly described. InF3 was obtained by fluora-
tion of In2O3 at 400 °C with NH4F and HF in a platinum
crucible. Then all fluoride components were mixed up and
heated in a dry box under argon atmosphere at 700 °C for
melting and 800 °C for finning. After this process the melt
was poured and cooled into a preheated brass mold. The
samples obtained have good optical quality, volumes of a
few cubic centimeters, and they are nonhygroscopic.
Optical absorption spectra in the 200–800 nm range were
obtained with a double-beam spectrophotometer, while the
infrared spectra up to 1.8 ␮m were measured with an optical
spectrum analyzer.
Continuous-wave upconversion fluorescence measure-
ments were performed using a diode laser emitting at 1.48
␮m as the excitation source. The laser beam was chopped at
7 Hz and focused on the sample using a lens of 15 cm focal
length. The sample fluorescence was collected perpendicu-
larly to the direction of the incident beam and was dispersed
by a 0.5-m grating spectrometer. The signal was detected
using either a GaAs or a S1 photomultiplier, and it was sent
to a lock-in amplifier or a digital oscilloscope connected to a
personal computer for processing.
Among the many fluoride compositions discovered, it was
found recently that fluoroindate glasses may become one im-
portant material for photonics applications. The vitreous re-
gion in the system InF3-ZnF2-͑SrF2-BaF2͒ was established
few years ago,2 and it was later observed that the glasses can
be stabilized by addition of GaF3, GdF3, CaF2, and NaF.3
Previous works performed in this system include the study of
their vibrational spectra and structure4 and the spectroscopy
of samples doped with Nd3ϩ,5 Pr3ϩ,5 Eu3ϩ,6 and Gd3ϩ.6 Our
recent studies have shown that when doped with Er3ϩ, Nd3ϩ
,
and Pr3ϩ, the fluoroindate glass presents large efficiency as
upconverters from the infrared to visible7 and from orange to
blue and violet.8,9
Erbium ions are appealing for spectroscopic investigation
due to their energy level regular spacing which facilitates
frequency upconversion via energy transfer or multistep
pump absorption using a single excitation wavelength. Ac-
cordingly, studies of upconversion in Er3ϩ-doped fluoride,10
borate,11 tellurite and gallate,12 and fluorophosphate glasses13
have been reported. Presently, it is well known that the up-
conversion efficiency is larger for fluoride glasses because
the multiphonon emission rates are much lower than the rates
for the same levels of Er3ϩ in other glasses.1,14
For the low-temperature measurements, the samples were
mounted in a cold-finger Dewar with temperature measured
by a thermocouple embedded in the mounting bracket. The
temperature of the samples could be varied from 24 to 300 K
0163-1829/97/55͑10͒/6335͑8͒/$10.00
55
6335
© 1997 The American Physical Society

´
6336
MACIEL, de ARAUJO, MESSADDEQ, AND AEGERTER
55
served for the different concentrations used because the elec-
tronic transitions within the 4f shell are not very sensitive to
the crystalline field. The spectra obtained for the other
samples are similar, except for the band intensities, which
depend on the Er3ϩ concentration. The linewidths of all tran-
sitions are inhomogeneously broadened due to the site-to-site
variation of the crystalline field strength.
Figure 2 shows the upconversion spectrum of the sample
with xϭ3 under 5.6 mW excitation ͑ϳ170 W/cm2͒ at room
temperature. The diode laser excites resonantly the
4
4I15/2→ I13/2 transition and the observed emissions corre-
spond to radiative transitions in the erbium ions from the
excited states. The distinct emissions correspond to the tran-
2
4
2
4
sitions H9/2→ I15/2 ͑ϳ407 nm͒, H11/2→ I15/2 ͑ϳ530 nm͒,
4
4
4
4S3/2→ I15/2 ͑ϳ550 nm͒, F9/2→ I15/2 ͑ϳ670 nm͒,
4
4
4I9/2→ I15/2 ͑ϳ808 and ϳ827 nm͒, 4S3/2→ I13/2 ͑ϳ854 nm͒,
4
4
and I11/2→ I15/2 ͑ϳ980 nm͒. The green and red transitions
are readily visible by the naked eye. The spectra were ana-
lyzed with respect to their pump power dependence and tem-
poral behavior. To analyze the results we first note that for
unsaturated frequency upconversion the fluorescence signal
IS will be proportional to some power n of the excitation
intensity such that ISϰIn, where nϭ2,3,4, . . . is the number
of infrared photons absorbed per upconverted photon emit-
ted. The dependence of the fluorescence signal on the
1.48-␮m excitation intensity was such that 3.6ϽnϽ3.9 for
the emission at 407 nm, 2.4ϽnϽ2.7 for the emission at 550
and 670 nm, 1.7ϽnϽ2.0 for the emission at 808 and 827
nm, 2.4ϽnϽ3.3 for the emission at 854 nm, and 1.8ϽnϽ2.0
for the emission at 980 nm. The measurements were made
for the three concentrations prepared, and the data for one of
the samples are shown in Fig. 3. From the intensity depen-
dence observed and the wavelength of the emitted radiations,
we conclude that four laser photons are involved in the 1.48-
␮m–to–407-nm conversion, three laser photons participate
in the 1.48 ␮m-to-550-nm, 1.48-␮m-to-670-nm, and 1.48-
␮m-to-854-nm conversions, and two laser photons produce
the 1.48-␮m-to-808-nm, 1.48-␮m-to-827-nm, and 1.48-␮m-
to-980-nm frequency upconversions. The deviation from the
exact n values are due to strong absorption at 1.48 ␮m, the
absorption of the upconverted fluorescence, and because the
nonradiative decay from higher-lying states to fluorescent
states may also contribute to the intensities of the observed
spectral lines. The fluorescence line peaked at 980 nm was
the most intense, being Ϸ50 times more intense than the
transition at 550 nm. For an incident power of 5 mW, about
1 ␮W is converted into 980 nm emission. We also observed
that for the sample with xϭ3 the signal at 980 nm was twice
larger than for the sample with xϭ2 and 12 times larger than
for the sample with xϭ1.
FIG. 1. Room-temperature absorbance spectra in the visible ͑a͒
and in the near-infrared ͑b͒ regions ͑sample with xϭ3, thickness 2.5
mm͒.
with regulation of 0.1 K. For the high-temperature experi-
ments, the samples were warmed up to 448 K by a
temperature-controlled thermal plate with regulation of
Ϯ1 K.
Different processes may lead to the population of highly
excited Er3ϩ states after excitation in the near infrared
III. RESULTS AND DISCUSSION
4
(4I15/2→ I13/2).13,15–19 These processes rely either on multi-
Figure 1 shows the room-temperature absorption spectra
in the visible range ͑a͒ and in the near infrared ͑b͒, obtained
for one of the samples prepared. The broad features of sev-
eral angstroms bandwidth can be identified with the transi-
tions from the ground state ͑4I15/2͒ to the excited states of the
Er3ϩ ions. The bands observed at 274 and 256 nm are due to
electronic transitions in the Gd3ϩ ions present in the glass
matrix. No changes in the wavelengths of maxima were ob-
step excited state absorption ͑ESA͒ or energy transfer ͑ET͒
between Er3ϩ ion neighbors. The ET process, in which an
excited ion nonradiatively transfers its energy to an already
excited neighbor, is one of the most efficient mechanisms
and has been observed in a large number of systems includ-
ing fluoroindate glasses.8–14 This mechanism can arise from
electric multipole or exchange interactions, and its rate oc-
currence depends on the RE concentration due to the ion-ion

55
FREQUENCY UPCONVERSION IN Er3ϩ-DOPED . . .
6337
FIG. 2. Room-temperature upconversion fluorescence spectra ͑sample with xϭ3͒. The intensities of ͑b͒, ͑c͒, and ͑d͒ have been multiplied
by 50, 40, and 0.02, respectively.
separation. In the present case, we expect that ET is the
dominant process because of the large Er3ϩ concentration in
our samples and because the intermediate ESA step
detected using a fast digital oscilloscope. The time resolution
of the detection system was better than 1 ms, and the signal
corresponding to the various upconverted emissions grew to
their maximum value in ␶rϽ15 ms and decay in ␶dϽ12 ms.
In general, the rise and decay times ͑␶ and ␶ ͒ decrease for
2
(4I13/2→ H11/2) is a two-photon transition with small prob-
ability to occur due to the laser frequency detuning for inter-
mediate states and the weak laser intensity used. Therefore,
the most relevant pathway for upconversion initiates with the
r
d
increasing concentrations and for the range of Er3ϩ concen-
trations studied ␶ and ␶ change up to 50%. The results
r
d
4
4
transition I15/2→ I13/2. Afterwards, ET between excited
Er3ϩ ions at the 4I13/2 level will take one ion to the 4I9/2 level.
This step is followed by other successive transfer processes
from ions at the 4I13/2 state, which results in the excitation to
higher levels. After nonradiative decay to lower states, radia-
tive transitions to the ground state give rise to the observed
upconverted fluorescence.
obtained using a digital oscilloscope are indicated in Table I.
In order to understand the observed temporal behavior,
we first recall that the upconversion efficiency depends on
the probability of multistep excitation by ESA or by ET be-
tween adjacent excited ions, as well as the quantum effi-
ciency of the emitting level. By either process, the dynamics
of the upconversion signals depends on the lifetime of the
intermediate excited states involved. For the samples used,
Figure 4 shows the relevant energy levels for the 4f11
configuration of an Er3ϩ ion, together with two possible up-
conversion pathways and the observed fluorescence lines.
To characterize the temporal evolution of the upconverted
signal, another series of experiments was performed. The
laser beam was chopped at 7 Hz, and the fluorescence was
2
4
4
4
the lifetime of the states H9/2, S3/2, F9/2, and I11/2 were
reported in Ref. 18. The values obtained for the same
range of Er3ϩ concentrations were ␶͑2H9/2͒ϳ15–20 ␮s,
␶͑4S3/2͒ϳ84–573 ␮s, ␶͑4F9/2͒ϳ302–645 ␮s, ␶͑4I11/2
͒
ϳ10.6–9.4 ms, and ␶͑4I13/2͒ϳ10.3 ms. The lifetime of the

´
6338
MACIEL, de ARAUJO, MESSADDEQ, AND AEGERTER
55
FIG. 4. Simplified energy scheme of Er3ϩ ion and excitation
mechanisms. The curved arrows on the right stand for energy trans-
fer. The letters beside the straight arrows correspond to the spectral
lines A ͑530 nm͒, B ͑550 nm͒, C ͑670 nm͒, D1 ͑808 and 827 nm͒,
D2 ͑854 nm͒, E ͑980 nm͒, F ͑407 nm͒, and G ͑377 nm͒.
A simple description of the frequency upconversion dy-
namics can be obtained in terms of rate equations for the
level populations. The equations can be written observing the
following aspects: First, from the fact that the upconversion
fluorescence intensity at 808, 827, and 980 nm increases qua-
dratically with the laser intensity, we conclude that these
emissions are due to the energy transfer process described by
4
4I13/2ϩ4I13/2→ I9/2ϩ4I15/2ϩphonons, followed by a decay
4
to level I11/2
. The emissions observed are due to
4I9/2→ I15/2 ͑808 and 827 nm͒ and I11/2→ I15/2 ͑980 nm͒.
On the other hand, the cubic dependence of the fluorescence
signals with laser intensity at 530, 550, 670, and 854 nm
indicates that they are due to an energy transfer redistribution
4
4
4
FIG. 3. Excitation intensity dependence of the upconversion
according to 4I13/2ϩ4I13/2ϩ4I13/2→ H11/2ϩ4I15/2ϩ4I15/2
2
fluorescence ͑sample with xϭ2 at room temperature͒. Straight lines
with different slopes n are obtained for each wavelength: ͑a͒ nϭ3.6
͑407 nm͒, nϭ2.7 ͑550 nm͒, nϭ2.4 ͑670 nm͒; ͑b͒ nϭ1.8 ͑808 nm͒,
nϭ1.8 ͑827 nm͒, nϭ3.3 ͑854 nm͒, nϭ1.9 ͑980 nm͒. The intensity
corresponding to the 980-nm line in ͑b͒ was scaled to be shown
with the other infrared fluorescence lines observed.
4
ϩphonons, followed by nonradiative decay to levels S3/2
and F7/2. The lines at 407 nm (2H9/2→ I15/2) and 377 nm
4
4
(4G9/2,4G11/2→ I15/2) are due to energy transfer among four
4
Er3ϩ ions in such a way that three ions deliver its energy to
the fourth ion that is promoted to the states 4G9/2, 4G11/2, and
2H9/2
.
In order to estimate the energy transfer rates between Er3ϩ
ions at room temperature, we used the following six-level
rate equation system:
4
state I9/2 was not measured, but on the basis of the known
results for other host materials, we expect that its magnitude
is Ϸ50 ␮s. The lifetimes are mainly determined by the mul-
tiphonon relaxation rates, which are small because of the
small phonon energies associated with the fluoroindate
matrix.4,10 Thus, considering that the states 4I9/2 and 4I11/2 are
likely to participate in the upconversion processes, we con-
clude that the large values observed for ␶ and ␶ provide
dn1
ϭϪR12n1ϩW2n22ϩ2W3n23ϩ␥21n2
dt
r
d
ϩ␥31n3ϩ␥41n4ϩ␥51n5ϩ␥61n6 ,
͑1͒
favorable evidence for the relevance of the ET mechanism.

55
FREQUENCY UPCONVERSION IN Er3ϩ-DOPED . . .
6339
TABLE I. Rise and decay times measured for the frequency upconversion signals at room temperature.
xϭ3
xϭ2
xϭ1
␭ ͑nm͒
␶
͑ms͒
␶
͑ms͒
␶
͑ms͒
␶
͑ms͒
␶
͑ms͒
␶
͑ms͒
decay
rise
decay
rise
decay
rise
407
550
670
808
827
854
980
9.9
3.6
4.5
4.2
4.3
4.7
10.9
9.9
9.6
6.8
6.7
3.9
4.9
4.7
4.9
4.9
11.5
4.1
5.6
5.1
5.4
5.1
8.1
7.7
5.2
5.7
7.8
13.0
11.5
7.7
7.2
15.0
15.7
5.6
11.0
10.6
12.9
5.8
11.3
6.0
11.5
11.9
measurements.10 The actual values used are depicted in
Table II. The software package MATHEMATICA was used to
solve numerically the system of equations described above
and to compare the results with the temporal behavior for
xϭ1 and 3. For a given value of W2 and W3, the theoretical
dynamics of population was found in such a way that the
fitting with the experimental data for every transition of our
system could be obtained with the same set of parameters.
The experimental rise times were very sensitive to the values
of W2 and W3, while the decay times were dependent almost
exclusively on the level lifetimes. The results obtained for
W2 and W3 are depicted in Table II. The values of W2 ͑W3͒
in Table III were obtained from W2 ͑W3͒ divided by the
concentration ͑squared concentration͒ of erbium ions, in or-
der to compare with values reported in the literature. Note
that W2 and W3 have the same order of magnitude than most
␥ij . Figure 5 shows the experimental results together with
the theoretical curves for samples with xϭ1 and 3.
dn2
dt
ϭR12n1Ϫ2W2n22Ϫ3W3n23Ϫ␥21n2
ϩ␥32n3ϩ␥42n4ϩ␥52n5ϩ␥62n6 ,
͑2͒
dn3
dt
ϭϪ␥3n3ϩ␥43n4ϩ␥53n5ϩ␥63n6 ,
͑3͒
͑4͒
͑5͒
͑6͒
dn4
ϭW2n22Ϫ␥4n4ϩ␥54n5ϩ␥64n6 ,
dt
dn5
ϭϪ␥5n5ϩ␥65n6 ,
dt
dn6
ϭW3n32Ϫ␥6n6 ,
dt
The temperature variation of the upconversion fluores-
cence was studied from 24 to 448 K for one of the samples
͑xϭ3͒, and the results are presented below. In Fig. 6͑a͒ it is
shown that the integrated fluorescence intensity at 407, 550,
4
4
4
where 1 stands for level I15/2, 2 for I13/2, 3 for I11/2, 4 for
4I9/2, 5 for F9/2, and 6 for S3/2. Although energy transfer
4
4
among triads at level 4I13/2, may promote one ion to the level
4H11/2, fast nonradiative decay ͑10 ns͒ populates the level
4S3/2. Thus we label in our model level 6 as the state S3/2
.
4
TABLE II. Values of decay rates used and fitting parameters
͑W2 and W3͒.
The contribution due to the population in level 4H9/2 was not
considered because of the small intensity of the blue fluores-
cence which originates from that level. We also did not con-
sider the nonresonant cross relaxation process
xϭ1
xϭ3
␥
␥
␥
␥
␥
␥
␥
␥
␥
␥
␥
␥
␥
␥
␥
(sϪ1
(sϪ1
(sϪ1
(sϪ1
(sϪ1
(sϪ1
(sϪ1
(sϪ1
(sϪ1
(sϪ1
(sϪ1
(sϪ1
(sϪ1
(sϪ1
(sϪ1
͒
͒
͒
͒
͒
͒
͒
͒
͒
͒
͒
͒
͒
͒
͒
100
79
13
115.3
95.76
13.62
116.72
35.70
3ϫ105
1115
206.41
43.60
3ϫ103
836.62
223.1
74.14
28.01
400
21
31
32
41
42
43
51
52
53
54
61
62
63
64
65
͑4I11/2ϩ4I13/2→ F9/2ϩ4I15/2ϩphonons͒ because the contri-
4
bution of this channel is assumed to be small in comparison
with the resonant ET process between three ions. In the
above rate equations, R12 is the pump rate which is given by
␴I/h␯ where ␴ϭ5.49ϫ10Ϫ25 m2 is the absorption cross sec-
tion of the transition, Iϭ6.37ϫ105 W/m2 is the laser inten-
sity, and h␯ϭ1.34ϫ10Ϫ19 J is the photon energy. It is as-
sumed that incident pump photons are resonant only with
transitions from the ground state to the first excited state
͑4I13/2͒. The quadratic terms ͑W2n 22 and Ϫ2W2n 22͒ and the
cubic terms ͑W3n 23, 2W3n 23, and Ϫ3W3n 32͒ describe the en-
ergy transfer between two and three atoms, respectively. W2
and W3 are the pair and trio energy transfer rates between
erbium ions, which are adjustable parameters in our model.
The relaxation rate ␥ij , i, jϭ1,2,3,..., is the radiative plus
nonradiative decay rate from level i to level j, and ␥i is the
total decay rate from level i. The radiative decays are ob-
tained using Judd-Ofelt theory,20 and the nonradiative relax-
ation rates were obtained from fluorozirconate glass
159.3
29.3
3ϫ105
1307.2
243
36.9
3ϫ103
674.1
180.1
59.8
22.1
400
R12 (sϪ1
͒
2.6
800
300
2.6
1000
500
W2 (sϪ1
W3 (sϪ1
͒
͒

´
6340
MACIEL, de ARAUJO, MESSADDEQ, AND AEGERTER
55
TABLE III. Energy-transfer-rate values for Er3ϩ in different hosts.
W2 ͑cm3/s͒
W3 ͑cm6/s͒
Host matrix
Concentration
Reference
1.5ϫ10Ϫ17
1ϫ10Ϫ16
YAG
silica fiber
YAlO3
16.6%
600 ppm
10%
21
22
23
2ϫ10Ϫ17
3.51–1.47ϫ10Ϫ18
5.77–1.08ϫ10Ϫ39
fluoroindate glass
1%–3%
Present work
and 670 nm decreases with increasing temperature in the
range from room temperature to 433 K. We note that in this
temperature interval the line at 670 nm shows a slower de-
crease of population with increasing temperature when com-
pared with the lines at 550 and 407 nm. When the tempera-
ture of the sample is lowered to below 300 K, the
upconverted fluorescence intensity increases by one order of
magnitude, and another emission peaked at 377 nm, corre-
4
sponding to 4G11/2→ I15/2 transition becomes detectable, but
its intensity is Ϸ104 times weaker than the 550 nm emission,
at 24 K. The relative intensities of the emission lines are
changed with respect to room temperature. For example, the
4
green line emitted from state S3/2 ͑ϳ550 nm͒ becomes so
strong that the windows of the Dewar are illuminated when
the room lights are off. On the other hand, the transition
4
2H11/2→ I15/2 at ϳ530 nm was not observed for tempera-
tures below 160 K. The results obtained are summarized in
Fig. 6͑b͒, which shows the temperature dependence of the
frequency upconversion signals at ϳ377, ϳ407, ϳ550, and
ϳ670 nm. For all lines the ratio between the intensity for a
given temperature with respect to the intensity at 295 K de-
creased when the temperature is increased above 50 K. How-
ever, below 50 K an increase of the signal intensity for all
the lines in the range 24–50 K is observed. This behavior at
low temperature indicates that population thermalization
does not occur below 50 K. We also observe that for tem-
peratures above 50 K the population temperature behavior is
similar to the one described in Fig. 6͑a͒. In Fig. 6͑c͒ we show
the ratio between the integrated intensities of light emitted at
670 and 550 nm. It can be observed that above 230 K the red
emission becomes stronger because of the significative mul-
4
4
tiphonon relaxation from the S3/2 state to the state F9/2
.
Another interesting result is that the relative intensity be-
tween the two lines at ϳ530 and ϳ550 nm changes with
temperature, which reveals the efficient thermal coupling be-
2
4
tween the excited states H11/2 and S3/2. Because of the
small energy gap between those levels ͑ϳ750 cmϪ1͒, the
2
4
state H11/2 may be populated from S3/2 by thermal excita-
tion and a quasithermal equilibrium occurs between the two
levels. This temperature behavior is consistent with previous
reports for fluoride glasses,10 but in the present case the fre-
quency upconversion process is so efficient that the fluoroin-
date glass could be used to build a sensitive temperature
sensor.15
IV. CONCLUSION
We have observed efficient frequency upconversion from
infrared to visible and to ultraviolet in Er3ϩ-doped fluoroin-
date glass. The analysis of the emitted radiation indicates
that the dominant mechanism involved in the upconversion
process is the energy transfer among excited erbium ions.
The large energy transfer rates observed are probably due to
dipole-dipole interactions, but in order to extend our knowl-
edge regarding the nature of the interactions among Er3ϩ
FIG. 5. Temporal behavior of the upconversion signal ͑a͒ emis-
sion at 808 nm ͑xϭ1͒ and ͑b͒ emission at 550 nm ͑xϭ3͒.

55
FREQUENCY UPCONVERSION IN Er3ϩ-DOPED . . .
6341
FIG. 6. Temperature dependence of the frequency upconversion integrated intensity at 407, 550, and 670 nm ͑a͒ for temperatures above
298 K and ͑b͒ for temperatures up to 298 K. The inset shows the behavior of the anti-Stokes signal at 377 nm. ͑c͒ Dependence of the
frequency upconversion integrated intensity ratio between the lines at 550 and 670 nm as a function of temperature ͑sample with xϭ3͒.
´
logico ͑CNPq͒, Financiadora Nacional de Estudos e Projetos
ions in this material, further investigation will be undertaken
in the future.
˜
˜
͑FINEP͒, Fundac¸ao Coordenac¸ao de Aperfeic¸oamento de
˜
Pessoal de Ensino Superior ͑CAPES͒, and Fundac¸ao de Am-
ACKNOWLEDGMENTS
`
ˆ
paro a Ciencia e Tecnologia ͑FACEPE͒. We thank also B. J.
P. da Silva for polishing the samples and TELEBRAS for the
loan of the diode laser used.
This work was supported by the Brazilian Agencies Con-
selho Nacional de Desenvolvimento Cientıfico e Tecno-
´
1 See, for example, Rare-Earth Doped Fiber Lasers and Amplifiers,
edited by M. J. Digonnet ͑Dekker, New York, 1993͒, and refer-
ences therein.
4 R. M. Almeida, J. C. Pereira, Y. Messaddeq, and M. A. Aegerter,
J. Non-Cryst. Solids 161, 105 ͑1993͒.
5 C. X. Cardoso, Y. Messaddeq, L. A. Nunes, and M. A. Aegerter,
J. Non-Cryst. Solids 161, 277 ͑1993͒.
2 Y. Messaddeq and M. Poulain, Mater. Sci. Forum 67–68, 161
͑1989͒.
6 S. J. L. Ribeiro, R. E. O. Diniz, Y. Messaddeq, L. A. O. Nunes,
3 Y. Messaddeq, A. Delben, M. A. Aegerter, A. Soufiane, and M.
Poulain, J. Non-Cryst. Solids 161, 210 ͑1993͒; J. Mater. Res. 8,
885 ͑1993͒.
and M. A. Aegerter, Chem. Phys. Lett. 220, 214 ͑1994͒.
7
´
Cid B. de Araujo, L. de S. Menezes, G. S. Maciel, L. H. Acioli,
A. S. L. Gomes, Y. Messaddeq, A. Florez, and M. A. Aegerter,

´
6342
MACIEL, de ARAUJO, MESSADDEQ, AND AEGERTER
55
´
Appl. Phys. Lett. 68, 602 ͑1996͒.
Araujo, Y. Messaddeq, A. Florez, and M. A. Aegerter, IEEE
Photonics Technol. Lett. 7, 1474 ͑1995͒.
8
´
´
L. E. E. de Araujo, A. S. L. Gomes, Cid B. de Araujo, Y. Mes-
saddeq, A. Florez, and M. A. Aegerter, Phys. Rev. B 50, 16 219
͑1994͒.
16 B. R. Reddy and P. Venkateswarlu, Appl. Phys. Lett. 64, 1327
͑1994͒.
9
17
´
¨
L. de S. Menezes, Cid B. de Araujo, Y. Messaddeq, and M. A.
M. P. Hehlen, G. Frei, and F. U. Gudel, Phys. Rev. B 50, 16 264
Aegerter, J. Non-Cryst. Solids ͑to be published͒.
͑1994͒.
10 D. C. Yeh, W. A. Sibley, M. Suscavage, and M. G. Drexhage, J.
Appl. Phys. 62, 266 ͑1987͒; M. D. Shinn, W. A. Sibley, M. G.
Drexhage, and R. N. Brown, Phys. Rev. B 27, 6635 ͑1983͒; K.
Hirao, S. Todoroki, and N. Soga, J. Non-Cryst. Solids 43, 40
͑1992͒.
18 R. Reiche, L. A. O. Nunes, C. C. Carvalho, Y. Messaddeq, and
M. A. Aegerter, Solid State Commun. 85, 773 ͑1993͒; T. Ca-
tunda, L. A. O. Nunes, A. Florez, Y. Messaddeq, and M. A.
Aegerter, Phys. Rev. B 53, 6065 ͑1996͒.
19 P. Xie and S. C. Rand, Appl. Phys. Lett. 63, 3125 ͑1993͒.
20 B. R. Judd, Phys. Rev. 127, 750 ͑1962͒; G. S. Ofelt, J. Chem.
Phys. 37, 511 ͑1962͒.
11 S. Tanabe, T. Ohyagi, N. Soga, and T. Hamada, Phys. Rev. B 46,
3305 ͑1992͒.
12 S. Tanabe, K. Hirao, and N. Soga, J. Non-Cryst. Solids 122, 79
͑1990͒.
21 W. Q. Shi, M. Bass, and M. Birnbaum, J. Opt. Soc. Am. B 7,
1456 ͑1990͒.
13 S. Tanabe, S. Yoshii, K. Hirao, and N. Soga, Phys. Rev. B 45,
4620 ͑1992͒.
22 P. Blixt, J. Niisson, T. Carlnas, and J. B. Jaskorzynska, IEEE
Photonics Technol. Lett. 3, 996 ͑1991͒.
14 X. Zou and T. Izumitani, J. Non-Cryst. Solids 162, 68 ͑1993͒.
15 G. S. Maciel, L. de S. Menezes, A. S. L. Gomes, Cid B. de
23 D. J. Simkin, J. A. Koningstein, P. Myslinski, S. A. Boothroyd,
and J. Chrostowski, J. Appl. Phys. 73, 8046 ͑1993͒.