Delayed release of Li atoms from laser ablated lithium niobate

Article abstract of DOI:10.1063/1.125847

The present vapor-phase optical (atomic) absorption measurements study the escape dynamics of Li atoms from a LiNbO₃ target surface upon laser ablation in vacuum. The objective is to understand the low-Li content of LiNbO₃ films prepared by pulsed laser deposition. A primary result is a delayed release of Li atoms, 2-20 μs after the laser pulse; they eject with a velocity of 6 × 10⁵ cms^-1, which is consistent with an electronic ejection mechanism. The long emission period means there are almost no intraplume Li collisions in the gas phase and no forward focusing of the delayed released atoms. This appears to explain the low-Li content usually found in films grown in the normal direction.

Full text of DOI:10.1063/1.125847

Delayed release of Li atoms from laser ablated lithium niobate
J. A. Chaos, R. W. Dreyfus, A. Perea, R. Serna, J. Gonzalo, and C. N. Afonso
Citation: Applied Physics Letters 76, 649 (2000); doi: 10.1063/1.125847
View online: http://dx.doi.org/10.1063/1.125847
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/76/5?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Imaging the ultrafast Kerr effect, free carrier generation, relaxation and ablation dynamics of Lithium Niobate
irradiated with femtosecond laser pulses
J. Appl. Phys. 116, 113502 (2014); 10.1063/1.4895833
Ferroelectric, piezoelectric, and leakage current properties of ( K 0.48 Na 0.48 Li 0.04 ) ( Nb 0.775 Ta 0.225 ) O 3
thin films grown by pulsed laser deposition
Appl. Phys. Lett. 98, 022902 (2011); 10.1063/1.3535608
Piezoresponse and ferroelectric properties of lead-free [ Bi 0.5 ( Na 0.7 K 0.2 Li 0.1 ) 0.5 ] Ti O 3 thin films by
pulsed laser deposition
Appl. Phys. Lett. 92, 222909 (2008); 10.1063/1.2938364
Cylindroid rigid-wall simulation of the influence of gas pressure in pulsed laser deposition of LiNbO 3 films
Appl. Phys. Lett. 82, 619 (2003); 10.1063/1.1540217
Concentration of Li atoms in plasmas produced from laser ablation of LiNbO 3
J. Appl. Phys. 90, 599 (2001); 10.1063/1.1381553
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.113.111.210 On: Sun, 21 Dec 2014 19:59:29

APPLIED PHYSICS LETTERS
VOLUME 76, NUMBER 5
31 JANUARY 2000
Delayed release of Li atoms from laser ablated lithium niobate
J. A. Chaos, R. W. Dreyfus,a) A. Perea, R. Serna, J. Gonzalo, and C. N. Afonsob)
Instituto de Optica, CSIC, Serrano 121, 28006 Madrid, Spain
͑
Received 28 May 1999; accepted for publication 6 December 1999͒
The present vapor-phase optical ͑atomic͒ absorption measurements study the escape dynamics of Li
atoms from a LiNbO target surface upon laser ablation in vacuum. The objective is to understand
3
the low-Li content of LiNbO films prepared by pulsed laser deposition. A primary result is a
3
5
delayed release of Li atoms, 2–20 ␮s after the laser pulse; they eject with a velocity of 6ϫ10
cm s , which is consistent with an electronic ejection mechanism. The long emission period means
Ϫ1
there are almost no intraplume Li collisions in the gas phase and no forward focusing of the delayed
released atoms. This appears to explain the low-Li content usually found in films grown in the
normal direction. © 2000 American Institute of Physics. ͓S0003-6951͑00͒04605-2͔
Lithium niobate crystals are well known to be of electro-
optical, piezoelectric, and acousto-optic usefulness. For the
development of compact waveguide devices for these appli-
data reported elsewhere for LiNbO .7 Even allowing for
3
some overestimation due to the finite 193 nm absorption
length, this result for ⌬T shows that the fluence used in this
experiment should be enough to melt and even vaporize
LiNbO3 .
cations, high-quality LiNbO films are required. Although
3
there have been many attempts to produce such films by
means of several techniques, limited success has been
achieved primarily because of the low-Li content. Pulsed la-
ser deposition ͑PLD͒, has been proven to have significant
success in growing complex oxides with similar perovskite-
like structure. Nevertheless, Li losses are still a problem in
The experiment uses a lithium hollow-cathode lamp to
measure the Li atomic absorption on the 670.8 nm unre-
2
o
/2
2
solved resonance doublet (2p P1 , 3/2� 2s S1/2). The ad-
vantages of this source are: cw, economical, always at the
correct wavelength, and a very narrow linewidth; the disad-
vantage is the low intensity of the source, sometimes being
overwhelmed by the plume light. The beam from the lamp
probes a cylindrical volume approximately 4 mm in diameter
and centered at a distance in the 7–25 mm interval in front of
the target surface and with the axis parallel to it. This mini-
mum distance has been selected since it was the distance at
which the contribution of the light emitted by the plume in
our experimental conditions was smaller than the light of the
Li lamp, and thus could be reliably substracted. A ϳ1:2 im-
age of the beam is formed into the 150 ␮m entrance slit of a
3/4 m Spex 1702 monochromator tuned to the resonance
wavelength to eliminate light from other lines. The probe
light is finally detected by a photomultiplier ͑Hammamatsu
R562͒ connected to a 100 MHz digital oscilloscope ͑Tek-
tronik, TDS 320͒ with a 1 k⍀ load resistor ͑to time average
photon statistics͒. The rise time of the detection system was
determined by analyzing the light emitted by the plume in
the neighborhood to the target and found to be 300 ns.
For each distance, several absorption transients were av-
eraged ͑256 shots͒ to improve the signal-to-noise ratio, spe-
cially at the longer distances, and two sets of transients were
always recorded: ͑i͒ transient Ioff(t) that corresponds to the
plasma light collected by the photomultiplier when the laser
fires but the lamp is off; and ͑ii͒ transient Ion(t) that corre-
sponds to the light collected when the lamp is on and the
laser fires. These records were then subtracted ͑to improve
the signal-to-noise ratio and to reduce the contribution of
plasma light if it were necessary͒ and the absorbance was
finally computed by comparing I(t)ϭI (t)ϪI (t) to the
PLD films of LiNbO , one of the best results reported so far
3
1
,2
being obtained by using Li-enriched targets. Furthermore,
low-Li content results in films with multiphases, strains, and
cracks. Some of the mechanisms proposed to explain the
2
,3
1
,4
loss of lithium include scattering by the ambient gas or
resputtering of the Li from the substrate.1
,5
The aim of this work is to study the dynamics of free-Li
atoms in the laser ablated LiNbO plume to explore the pro-
3
cesses which might lead to the low-Li content usually re-
ported in the films grown by the PLD technique. We use
optical ͑atomic͒ absorption as a measurement technique. Ab-
lation is produced by an ArF laser beam focused onto a
LiNbO single crystal at an incident angle of 45° from the
3
normal to the target. Two LiNbO crystals, cut one with the
3
c axis forming an angle of 14° with the normal to the surface
and one with the c axis in the surface plane, were studied.
The targets are mounted in a rotating holder and placed in a
vacuum chamber. The experiments are performed in vacuum
Ϫ7
at a background pressure of 6ϫ10 mbar. The beam depos-
Ϫ2
its Iϭ0.06 J cm into a 0.5ϫ4 mm ellipse; this low-energy
density has been chosen since it is very close to the plume
formation threshold and thus gives significant Li absorption
6
but little emission. A straightforward calculation gives a
temperature rise ⌬T in excess of 2000 K for the fluence
used. We have used the expression ⌬Tϭ(Dt/␳)1 (1
ϪR)2I/␬, where ␳ is the density, t is the ArF laser effective
pulse duration, D is the thermal diffusivity, R is the reflec-
tance, and ␬ is the thermal conductivity, and inserted the
/2
on
off
incident light Io ͑the lamp is on but the laser has not yet
a͒
Present address: 404 Adams St., Annapolis, MD 21403.
Author to whom correspondence should be addressed; electronic mail:
cnafonso@io.cfmac.csic.es
b͒
fired͒, the optical density being log͓I /I(t)͔.
o
A typical result of the optical-absorption measurements
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
003-6951/2000/76(5)/649/3/$17.00 649 © 2000 American Institute of Physics
30.113.111.210 On: Sun, 21 Dec 2014 19:59:29
0
1

650
Appl. Phys. Lett., Vol. 76, No. 5, 31 January 2000
Chaos et al.
FIG. 2. Optical density of the absorption of Li atoms recorded at three
distances from the target surface. Those recorded at 15 and 19 mm have
been normalized to the same intensity maximum as the one recorded at 11
mm.
FIG. 1. Optical density of the absorption of Li atoms recorded at ͑1͒ 11 mm,
2͒ 15 mm, ͑3͒ 19 mm, and ͑4͒ 23 mm from the target surface. The inset
͑
shows the distance at which the measurement is done as a function of the
time at which the maximum of the absorption occurs. The data were ob-
tained sequentially by increasing the distance ͑᭿͒ or alternatively by de-
creasing the distance (᭝,᭞). The dashed line is a linear fit of the experi-
mental data.
the Maxwellian type of function reported elsewhere10 that
fitted the FWHM of the experimental transient recorded at
the shortest distance measured ͑7 mm͒. Using the parameters
of such a function, we have then determined the FWHM
expected for the other distances studied if the evolution were
of Maxwellian type. Both the experimental and calculated
results are plotted in Fig. 3, where it is seen that the expected
broadening of the FWHM with a Maxwellian type of veloc-
ity distribution is much larger than that observed experimen-
tally. We can thus conclude that the absorption transients
relate to atoms which are ejected during a long time interval
with similar velocities, rather than to atoms released in a
short time interval with significantly different velocities.
The next step is to note the direct effect of this delayed
release. Gas-phase collisions are required to form a Knudsen
layer with its forward-directed collimation, and under typical
conditions this happens a few or tens of nanoseconds after
the laser pulse.10 The delayed and long-lasting Li desorption
will mean few or no gas-phase collisions for these delayed Li
atoms within the plume. A significant amount of Li atoms
thus should expand at large angles and in those experiments
is shown in Fig. 1 where the absorption of Li atoms recorded
at different distances d to the target is plotted. Similar results
were obtained in several sets of data collected, and indepen-
dently of the order followed, to record the absorption tran-
sients as a function of the distance. The results obtained us-
ing the two different targets were also similar, and thus only
the results for the one having the c axis forming an angle of
14° with the normal to the surface are shown. The transients
plotted in Fig. 1 evidence that the absorption of Li atoms
peaks at about 2–6 ␮s and lasts longer than 10 ␮s. The
intensity of absorption decreases with increasing distance, as
could be expected from the angular spreading of the Li at-
oms. The time at which the maximum occurs has been plot-
ted in the inset of Fig. 1 as a function of the distance to the
target. It is clearly seen that the experimental results follow a
linear relationship; this provides the expansion velocity of
the Li atoms. This velocity turns out to be quite high, 6
5
Ϫ1
ϫ10 cm s , which amounts to 1.3 eV, and is similar to
that reported earlier for the Li atoms generated immediately
after the laser pulse ͑hundreds of nanoseconds͒ and measured
dealing with LiNbO film growth, will contribute very little
3
to maintain the stoichiometry of Li along the direction per-
pendicular to the target.
From the point of view of velocities, direct thermal pro-
cesses do not seem to explain the origin of this delayed Li
8
through emission experiments. The long period together
with the high absorbance means a large flux of Li atoms.
This should represent a significant or even major fraction of
the free-Li atoms being generated. Evaluation of the Li time-
integrated optical density versus the volume of the target
removed reinforces this belief.
Both the emission and absorption transients have occa-
sionally been interpreted as time-of-flight velocity
9
,10
distributions.
If the transients primarily correspond to a
velocity distribution, they should broaden in time and shift
their maximum to longer times due to the spreading of the Li
atoms. Figure 2 shows some transients recorded in the 11–19
mm distance interval from the target and normalized to the
same intensity. It is clearly seen that all of them have very
similar shapes, and thus cannot be related to a velocity dis-
tribution. Furthermore, no good fit of the experimental tran-
sients could be made using the Maxwellian-type velocity dis-
FIG. 3. FWHM of the experimentally measured optical density transients as
a function of the distance from the target. The data were obtained sequen-
tially by increasing the distance ͑᭿͒ or alternatively by decreasing the dis-
tance (᭝,᭞) and they correspond to the same set of data whose time posi-
tion is included in the inset of Fig. 1. The dashed line corresponds to the
expected evolution for the FWHM of the velocity distribution assuming the
1
0
tributions widely reported in the literature. To prove further
that the transients are not consistent with a velocity distribu-
tion, we have measured the full width at half maximum
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
͑
FWHM͒ of the absorption transients. We have calculated
Maxwellian type of function described in Ref. 10.
1
30.113.111.210 On: Sun, 21 Dec 2014 19:59:29

Appl. Phys. Lett., Vol. 76, No. 5, 31 January 2000
Chaos et al.
651
release. First, one can calculate that the LiNbO target sur-
face has cooled to nearly ambient temperature after 10 or 20
plain the low-Li content usually reported in the films depos-
ited in the forward-directed direction.
3
␮
s, so direct vaporization cannot be responsible. Also for
This study was partially funded by CICYT ͑Spain͒ under
Project No. TIC96-0467. One of the authors ͑R.W.D.͒
wishes to acknowledge financial support from Fundacion
BBV. One of the Authors ͑J.G.͒ acknowledges a reincorpo-
ration contract from the Spanish Ministry of Education and
thermal emission of Li a direct approach to the velocities
would have given Ͻ10 cm s , in disagreement with the
velocity calculated from the data shown in Fig. 1 (Ͼ5
ϫ10 cm s ). Thus, the Li atoms’ kinetic energies are too
high for a thermal process. These energies are typical for
electrostatic, plasma, or color-center promoted emission.
5
Ϫ1
5
Ϫ1
Culture. Helpful discussions on LiNbO and the targets were
3
supplied by Dr. E. Dieguez ͑UAM, Madrid, Spain͒ and com-
puter solutions were kindly supplied by M. C. Morilla and J.
Siegel ͑Instituto de Optica͒.
Changes of stoichiometry in the ablated zone of LiNbO
3
crystals have often been reported,1
ute to defect-formation and thus to defect-promoted
release.
1–13
which might contrib-
1
4,15
It has also been reported that surface fracture can
1
Y. Shibata, K. Kaya, K. Akashi, M. Knai, T. Kawai, and S. Kawai, J.
promote the ejection of neutral atoms, although through a
thermal mechanism.16 In order to analyze the validity of
Appl. Phys. 77, 1498 ͑1995͒.
M. Haruna, H. Ishizuki, J. Tsutsumi, Y. Shimaoka, and H. Nishihara, Jpn.
2
these mechanisms, we also performed experiments accumu-
lating a number of pulses on the two studied targets without
rotating them. The crater produced has been observed by
optical microscopy, and although cracking and/or faceting
during recrystallization of the target was observed, they were
only observed for one of the studied target orientations. The
fact that the same Li dynamics were deduced from the ab-
sorption transients, suggests that the long-lasting delayed re-
lease of Li atoms is not related to the morphological changes
produced in the target surface upon ablation. This result
leaves color-center/defect-promoted ablation as the remain-
ing plausible explanation. The detailed mechanism is beyond
the scope of this work and requires further investigation.
In summary, a delayed and long-lasting release of fast Li
J. Appl. Phys., Part 1 34, 6085 ͑1995͒.
S. Bauer, Thesis, Universit a¨ t K o¨ ln, Mai 1995, JUEL-3056, ISSN 0944-
3
2952.
4
J. Gonzalo, C. N. Afonso, J. M. Ballesteros, A. Grossman, and C. Ortega,
J. Appl. Phys. 82, 3129 ͑1997͒.
S.-H. Lee, T. W. Noh, and J.-H. Lee, Appl. Phys. Lett. 68, 472 ͑1996͒.
For the formula, see H. S. Carslaw and J. C. Jaeger, Conduction of Heat in
Solids ͑Oxford University, New York, 1948͒, p. 56.
T. H. Lin and S. H. Lee, in Properties of Lithium Niobate, EMIS Data
Review Ser. 5 ͑Inspec, London, 1989͒, p. 56.
C. N. Afonso, F. Vega, J. Gonzalo, and C. Zaldo, Appl. Surf. Sci. 69, 149
͑1993͒.
D. B. Geohegan, in Pulsed Laser Deposition of Thin Films, edited by D.
B. Chrisey and G. K. Huber ͑Wiley, New York, 1996͒, Chap. 5, p. 115.
R. Kelly and R. W. Dreyfus, Surf. Sci. 198, 263 ͑1988͒.
M. Affatigato, K. Tang, R. F. Haglund, Jr., and C. H. Chen, Appl. Phys.
Lett. 65, 1751 ͑1994͒.
N. Omori and M. Inoue, Appl. Surf. Sci. 54, 232 ͑1992͒.
M. Sorescu, E. T. Knobbe, J. J. Martin, J. D. Barrie, and D. Barb, J. Mater.
Sci. 30, 5944 ͑1995͒.
M. A. Schildbach and A. V. Hamza, Phys. Rev. B 45, 6197 ͑1992͒.
N. Seifert, D. Liu, R. G. Albridge, A. V. Barnes, N. Tolk, W. Husinsky,
and G. Betz, Phys. Rev. B 46, 83 ͑1992͒.
R. L. Webb, L. C. Jensen, S. C. Langford, and J. T. Dickinson, J. Appl.
Phys. 74, 2323 ͑1993͒; J. T. Dickinson, L. C. Jensen, R. L. Webb, M. L.
Dawes, and S. C. Langford, ibid. 74, 3758 ͑1993͒.
5
6
7
8
9
10
11
12
atoms upon ablation of LiNbO crystals has been identified.
3
13
The high velocities of the atoms are consistent with an elec-
tronic mechanism. The delayed release means that the ma-
jority of the Li atoms do not suffer gas-phase collisions
within the plume and are thus not forward directed as the
1
1
4
5
16
other species and only slightly contribute to LiNbO film
3
growth. These different dynamics for Li atoms thus can ex-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
30.113.111.210 On: Sun, 21 Dec 2014 19:59:29
1