Home Our contacts
Home About SPTS Products Services How it works Contacts
Page 1
Green-induced infrared absorption in MgO doped LiNbO3
Y. Furukawa
a)
OXIDE Corporation, 9633 Kobuchisawa, Kitakoma, Yamanashi 408-0044, Japan
K. Kitamura
National Institute for Research in Inorganic Materials, 1-1 Namiki, Tsukuba 305-0044, Japan
A. Alexandrovski, R. K. Route, and M. M. Fejer
E.L. Ginzton Laboratory, Stanford University, Stanford, California 94305
G. Foulon
Crystal Technology Inc., 1040 East Meadow CIRcle, Palo Alto, California 94303
Received 25 September 2000; accepted for publication 2 February 2001
Green-induced infrared absorption GRIIRA was investigated by a photothermal technique for
undoped and Mg-doped LiNbO3
crystals that have different Li/Nb ratios. Threshold effect on
GRIIRA was found, threshold MgO concentrations being the same for GRIIRA and photorefraction.
We suggest that GRIIRA is associated with the formation of the small polaron that is located on Nb
antisite defect. The remarkable decrease of GRIIRA in Mg:LiNbO3
can then be attributed to the
elimination of this intrinsic defect, Nb in Li, following the incorporation of Mg on Li sites. For
nonlinear optical applications, LiNbO3
doped with MgO at concentrations over threshold has a
combined advantage of having almost no GRIIRA and photorefraction. © 2001 American Institute
of Physics. DOI: 10.1063/1.1359137
In recent years there has been increasing interest in the
use of quasi-phase-matched QPM nonlinear crystals
1
for a
variety of frequency conversion applications. Periodically
poled lithium niobate PPLN has been demonstrated to give
efficient second-harmonic generation
2
and optical parametric
oscillation
3
in both the cw and the Q-switched regimes.
However, the performance of high power operating PPLN-
SHG devices has been limited by material issues such as
photorefractive beam distortion and green-induced infrared
absorption GRIIRA .
4
It has been demonstrated that the
former problem can be solved by either high temperature
device operation or MgO doping to the LiNbO3
crystal, how-
ever, the latter problem still remains to be investigated. De-
fect models of the phenomena of green-induced infrared ab-
sorption have not been developed. Recently we reported that
photorefraction in LiNbO3
was completely eliminated by the
doping of small amounts of MgO in crystals with near-
stoichiometric composition.
5
These notable changes in mate-
rial properties are strongly related to the elimination of Nb
antisite defects (Nb
Li
) by the substitution of Mg on Li sites.
In this letter, we investigate the influence of MgO dop-
ing on the GRIIRA properties in LiNbO3
, and demonstrate
that LiNbO3
crystals doped with MgO show remarkably
lower GRIIRA than undoped LiNbO3
crystals.
LiNbO3
single crystals grown by two different methods
were used in this study. They are denoted CLN when theIR
stoichiometry, expressed as the ratio c
Li
=[Li] /( [Li]+
[Nb] ), has a congruent-composition value of 48.4% and
SLN when it deviates from this value in the direction of the
stoichiometric composition of 50%. crystals denoted as SLN
and Mg:SLN were grown as described in earlier reports
6
by
use of a top-seeded solution growth technique from a sto-
ichiometric melt with the addition of 11 mol % of K
2
O as
flux and of MgO as dopant. crystals denoted as CLN and
Mg:CLN were grown by the conventional Czochralski
method from congruent melt compositions. Chemical com-
position, Curie temperature, OH absorption wave number,
and photorefractive damage threshold of the LiNbO3
single
crystals used in this study are summarized in Table I.
7
The
corresponding chemical formulas were characterized by
chemical analysis as follows: The crushed samples were dis-
solved with HNO
3
–HF solution in a closed Teflon vessel at
150 °C for one night. After being filtered for the separation
of precipitated MgF
2
, the yielded solution was passed
through as SA-a anion exchanger for further separation of Li
and Nb ions. The Li and Mg content were analyzed by an
inductively coupled plasma atomic emission spectrometry
ICP-AES . The Nb ions eliminated from the resin by
HCl–HF solution were precipitated by cuperon, and the pre-
cipitate was dried and incinerated. The weight of the yielded
Nb
2
O
5
gave the Nb content. The accuracy for Li / Nb cat-
ion ratios in Mg:SLN and Mg:CLN was 0.6% while that for
MgO content amounted to 0.8%. As shown in Table I, it is
clear that the SLN crystals contain lower intrinsic defect den-
sities than conventional CLN crystals, but they still contain
Nb
Li
at levels of thousands of ppm. On the other hand, ex-
trinsic defects such as Fe were determined by ICP analysis to
be in the ppm levels. In Table I, optical damage threshold
refers to the onset of optical damage. The high MgO concen-
tration samples, according to the data presented, displayed
no optical damage, therefore the optical damage threshold
was never reached. The high MgO concentration samples
showed a shift in the OH absorption band from 3466 to 3485
to 3552 cm
-1
which suggests that at those doping levels, the
excess Nb
5+
on Li sites has been completely replaced by
Mg
2+
.
5
a
APPLIED PHYSICS LETTERS
VOLUME 78, NUMBER 14
2 APRIL 2001
1970
© 2001 American Institute of Physics

Page 2
The technique for absorption measurements used in this
work was photothermal common-path interferometry
8
which
has the sensitivity of 0.1 ppm/cm. We measured absorption
of an infrared pump beam wavelength 1064 nm with the
power density of 21 kW/cm
2
, which was focused into 0.5
mm thick crystals. Both green and infrared light were propa-
gated along the z axis in order to separate refractive index
change caused by thermal effect from photorefractive effect.
A green beam, with a wavelength of 532 nm and with the
power density in the range 0.03–3.6 kW/cm
2
, was opened
and closed manually. Green and IR beams were aligned to
overlap inside the crystal so that the change of IR absorption
in the presence of green light could be monitored. Values for
photorefraction in the presence of green light were estimated
under the same conditions as for the GRIIRA measurements.
These values are estimated from the basic feature of the pho-
tothermal system, which detects any phase distortion of the
probe beam caused by temperature change or photorefrac-
tion. Note that for the specific configuration used light
propagating along the optic axis, thin crystal the observed
index change is approximately two orders of magnitude
lower than for the conventional geometry with extraordinary
polarized light propagating in the dIRection normal to the
optic axis.
Figure 1 shows an example of green-induced infrared
absorption and photorefraction as a function of time for an
undoped, nearly stoichiometric LiNbO3
crystal ( Li / Nb
0.988). As seen in this figure, undoped SLN shows small
initial IR absorption (0.0017 cm
-1
) and a factor of five in-
crease in IR absorption in the presence of green light. The
green induced IR absorption observed here recovered to its
original value when the green pump beam was turned off.
Figure 2 shows the effect of Li / Nb stoichiometry and
Mg doping on the GRIIRA in LiNbO3
. For low doping lev-
els GRIIRA increased with the increase in Mg concentration,
however, doping with Mg concentrations that exceeded cer-
tain threshold levels showed remarkable decreases of
GRIIRA in both SLN and CLN cases. Mg:SLN requIRes a
smaller Mg concentration than Mg:CLN for low GRIIRA
state to be reached. The data plotted by open circles and
open triangles in Fig. 2 show samples that exhibit no mea-
surable photorefractive damage at 532 nm to intensities of as
much as 8 MW/cm
2
in Table I.
Figure 3 shows photorefraction as a function of Li /
FIG. 1. Example of the change of infrared absorption and photorefraction by
green light irradiation in an undoped nearly stoichiometric LiNbO3
crystal
listed in Table I.
TABLE I. Chemical formula, Curie temperature, wave number of OH absorption, and photorefractive damage
threshold of samples.
Sample
Chemical formula
Curie temperature
°C
Position of
OH absorption
band (cm
1
)
Optical damage
threshold
a
(kW/cm
2
)
Nondoped CLN
Li
0.950 0.04
Nb
1.01
O
3
1145.0
3485
1
Mg30 CLN
Li
0.925
Mg
0.030 0.042
Nb
1.003
O
3
1188.5
3485
10
Mg45 CLN
Li
0.901
Mg
0.045 0.052
Nb
1.0018
O
3
1209.0
3485
75
Mg60 CLN
Li
0.886
Mg
0.060 0.055
Nb
0.9987
O
3
1220.0
3532
8000
b
Nondoped SLN
Li
0.990 0.008
Nb
1.002
O
3
1192.0
3466
0.1
Mg06 SLN
Li
0.9856
Mg
0.0072 0.0066
Nb
1.0006
O
3
1205.5
3466
10
Mg18 SLN
Li
0.972
Mg
0.019 0.011
Nb
0.9980
O
3
1211.0
3532
8000
b
Mg36 SLN
Li
0.957
Mg
0.034 0.14
Nb
0.9950
O
3
1210.5
3532
8000
b
a
Photorefractive damage threshold is defined as the cw green-light intensity where the transmitted laser beam is
distorted as a result of photorefraction after 10 min of IRradiation see Ref. 5 .
b
No photorefraction was observed at this intensity level.
FIG. 2. Green induced infrared absorption GRIIRA vs MgO concentration
in SLN and CLN crystals. Chemical formula and other material properties
of these samples are listed in Table I.
1971
Appl. Phys. Lett., Vol. 78, No. 14, 2 April 2001
Furukawa
et al.

Page 3
Nb and MgO concentrations. The refractive index changes
were measured along with the GRIIRA measurementss shown
in Fig. 2. The photorefraction in both SLN and CLN crystals
decreased monotonically with increase of MgO concentra-
tion. SLN crystals with high Li / Nb ratios show faster de-
creases of photorefraction than congruent crystals. The fea-
ture similar to the GRIIRA behavior is the threshold doping
level over which photorefraction disappears along with the
disappearance of GRIIRA.
The reasons for the threshold behavior of both photore-
fraction and GRIIRA are probably related. Moreover, in
Table I, the optical damage disappearance and the shift in the
OH absorption band from 3466 to 3485 to 3532 cm
-1
both
correspond to the same threshold MgO concentrations. The
comparison with the chemical formulas strongly suggests
that at these doping levels the excess Nb
5+
on Li sites has
been completely replaced by Mg
2+
.
5
We think that GRIIRA is associated with small polaron
absorption band that is centered in the near IR, much like it
occurs in reduced crystals though the absorption is several
orders of magnitude less for crystals measured in this work.
The small polaron is believed to be formed by electron
trapped by antisite Nb.
9,10
The results of an early work
11
suggest that at room temperature that this trap is unstable, the
decay time being in the range of milliseconds. This explains
the rapid response of induced absorption on the green light
power observed in Fig. 1. The significant decrease in GRI-
IRA over a threshold concentration can be attributed to the
complete elimination of intrinsic defects of Nb
Li
following
the incorporation of Mg on Li sites.
The increase of GRIIRA with doping in the region under
threshold needs some comments. Both the rate of excitation
of electrons to the conduction band and small polaron life-
time will contribute to the IR absorption along with Nb
Li
concentration. As to the concentration, it is clear that un-
doped SLN crystals contain lower intrinsic defect densities
than conventional CLN crystals, but they still contain Nb
Li
at
a level of thousands of ppm. Since SLN crystals show stron-
ger GRIIRA than CLN crystals the other factors should play
a very important role.
As for the excitation rate, it should be proportional to the
concentration of the corresponding defects responsible for
the green absorption. Fe impurity is known to contribute to
excitation of electrons to the conduction band. The concen-
tration of Fe was determined by chemical ICP analysis to
be in the ppm levels both in SLN and CLN. This number is
well below the concentration of Nb
Li
shallow traps, even in
SLN, so that it is likely that most excited electrons will be
trapped by Nb
Li
. This is why the dominant factor in
GRIIRA, for the same excitation rate, is probably small po-
laron lifetime. In this way it is possible that small polaron
lifetime increases dramatically near the threshold. This as-
sumption has to be checked experimentally.
In summary, we have investigated green-induced infra-
red absorption and photorefraction by means of the photo-
thermal common-path interferometry technique. We ob-
served that MgO doping had similar results for reducing
GRIIRA in LiNbO3
, namely that less MgO was necessary to
eliminate GRIIRA in stoichiometric LiNbO3
than in congru-
ent LiNbO3
. LiNbO3
crystals doped with MgO at levels ex-
ceeding threshold concentrations of 1 mol. % for SLN and 5
mol. % for CLN exhibit no measurable GRIIRA with an IR-
radiation up to 3.5 kW/cm
2
of green light. This remarkable
decrease of GRIIRA in Mg:LiNbO3
is attributed to the elimi-
nation of antisite intrinsic defects of Nb
Li
.
1
M. M. Fejer, G. A. Magel, D. H. Jundt, and R. L. Byer, IEEE J. Quantum
Electron. QE-28, 2631 1992 .
2
K. Mizuuchi, K. Yamamoto, and M. Kato, Electron. Lett. 25, 731 1997 .
3
L. E. Myers, G. D. Miller, R. C. Eckardt, M. M. Fejer, R. L. Byer, and W.
R. Bosenberg, Opt. Lett. 20, 52 1995 .
4
R. G. Batchko, G. D. Miller, A. Alexandrovski, M. M. Fejer, and R. L.
Byer, 1998 OSA Technical Digest Series Optical Society of America,
Washington, DC, 1998 , Vol. 12, pp. 75–76.
5
Y. Furukawa, K. Kitamura, S. Takekawa, K. Niwa, and H. Hatano, Opt.
Lett. 23, 1892 1998 .
6
K. Niwa, Y. Furukawa, K. Kitamura, S. Takekawa, and H. Hatano, J.
Cryst. Growth 208, 493 2000 .
7
Y. Furukawa, K. Kitamura, S. Takekawa, K. Niwa, Y. Yajima, N. Iyi, I.
Mnushkina, P. Guggenhein, and J. Martin, J. Cryst. Growth 211, 230
2000 .
8
A. Alexandrovski, G. Foulon, L. E. Myers, R. K. Route, and M. M. Fejer,
Proc. SPIE 3610, 44 1999 .
9
K. Buse, F. Jermann, and E. Krätzig, Opt. Mater. 4, 237 1995 .
10
F. Jermann, M. Simon, and E. Krätzig, J. Opt. Soc. Am. B 12, 2066
1995 .
11
E. G. Valyashko and V. A. Timoshenkov, Zh. Prikl. Spektrosk. 15, 1008
1971 .
FIG. 3. Photorefractive index change vs MgO concentration in SLN and
CLN crystals. Chemical formula and other material properties of these
samples are listed in Table I.
1972
Appl. Phys. Lett., Vol. 78, No. 14, 2 April 2001
Furukawa
et al.