czochralski growth , crystalline perfection and optical c...
TRANSCRIPT
Chapter – 4 LiNbO 3 & Zn:LiNbO 3
~ 87 ~
Chapter - 4
Czochralski growth, crystalline
perfection and optical characterization
of LiNbO3 and Zn:LiNbO3 NLO single
crystals
Abstract
The pure and Zn-doped (1 mol%) LiNbO3 single crystals have been
grown by Czochralski method. The structure of the crystals has been
confirmed and the strains in both the crystals have been evaluated by
powder X-ray diffraction. The high resolution multicrystal X -ray
diffractometer has been employed to record the rocking for (006)
planes to assess the crystalline perfection of the grown crystals . The
Fourier transform infrared and Raman spectra have been recorded for
investigation of vibrational modes, defects, Li-composition and
incorporation of protons (H+
) in crystals’ lattice . The UV-VIS-NIR
studies have been carried out to study the transparency, band gap and
defectiveness. The birefringence and refractive indices for 1064 and
532 nm wavelengths have been studied using prism coupler
spectrometer. The wavelengths dispersion , optical dielectric constant,
dispersion energy ‘Ed’ and average single oscillator energy ‘E o’ for
the crystals have been studied by the Ellipsometry.
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 88 ~
4.1 INTRODUCTION
The single crystals are backbone for almost all the advanced scientific and
technological developments. The modern science and technology demand fast data
processing and high density storage devices (Carina et al., 2002). The optical
holographic systems are proposed to be the best to overcome the high density data
handling difficulties (Heerden et al., 1963; Mark Haw, 2003). The oxide single
crystals play an unmatched role for their applications in lasers, nonlinear optical
devices employing for storing, retrieving and processing the information in the form
of photon at the speed of light. The Lithium Niobate (LiNbO3; LN), a nonlinear
optical (NLO) material, has been well proved as a potential and unique material for its
use in advanced photonic device applications like second-harmonic generation (SHG),
high energy lasers (for inertial confinement fusion reactions), electro-optical
switching, optical modulators, holographic data storage, acousto-optic, optical
communication and ferroelectric. Lithium Niobate, due to its role in widespread
photonics era is also termed as ‘silicon of photonics’. Zn doping removes the antisitic
Nb (NbLi4+
) and improves the optical damage resistance significantly and sufficed the
recording of holograms at high-speed with long life-time by resulting in high
photoconductivity and optical band gap of LN (Zhen et al., 2003; Volk et al., 1996).
The SHG efficiency of LN in crystals was found to be increased manifold with Zn-
doping (Nevado, R. & Lifante, 2001). Fabrication of LN waveguides with low
propagation loss and high photorefractive damage resistance has been reported by
diffusion of Zn (Li et al., 2001). The Enhanced SHG has also been observed in
periodically poled Zn-doped LN planar waveguide (Vincent et al., 2007). The Nd3+
co-doped LN:ZnO crystal exhibited continuous-wave laser action at 929 nm suitable
for generation of blue lasers (Jaque et al., 2004). Analogous to In and Na doping, Zn
doping makes LN suitable for ultraviolet recording grating application (Haijun Qiao
et al., 2004) and recently, Feifei, et al. (2010) have demonstrated ultraviolet
photorefraction at 325 nm in highly Zn-doped LN for its holographic applications in
UV region. The polarization independent signal (wavelength) conversion with
ultrafast modulation could be achieved in LN with Zn doping (Masaki et al., 2003).
The cited literature proves the importance of Zn doped LiNbO3 single crystals in
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 89 ~
photonics. However, the need of systematic study on the defect structure in a carefully
grown LN crystal due to doping particularly with Zn which influences the physical
properties motivated the present investigation which is elaborated in the following.
The properties of the doped single crystals depend on the fact that weather the dopants
are homogeneously distributed in the entire crystal or agglomerated into the clusters.
The homogeneous distribution of dopants in the lattice matrix of host crystal is being
decided by the concentration, size, ionic state of dopants and the lattice sites of
crystals occupied by the dopants. In LN, the dopants distribution is statistical up to
their certain concentration known as ‘threshold concentration’. Above this
concentration of dopants, high geometric strains develop in lattice and the lattice tries
to relax by overcoming the strains, during this process the dynamics of point defects
take place which lead to the agglomeration of point defects and dislocations, and this
ultimately results in the formation of structural grain boundaries (Bhagavannarayana
et al., 2009; Bhagavannarayana et al., 2005). The properties of single crystals are
anisotropic in nature, therefore in the presence of structural defects these get masked
or partially/completely deteriorate (Haixuan et al., 2009; Tsai et al., 2008; Dongfeng
& Xiangke, 2006; Bhagavannarayana et al., 2006; Bhagavannarayana & Kushwaha,
2010) and thereby may reduce the efficiency of device property. Therefore, for the
realization of full efficiency of the device, the crystal must be free from such defects
(Bhagavannarayana, Budakoti et al. 2005).
In the modern era of miniaturized technology of photonics it is very important
to grow the pure and doped bulk single crystals of LN and assess their crystalline
perfection, especially in the case of doped crystals. The single crystals of LN have
been grown by different techniques depending on their application. However, the
Czochralski (CZ) technique has been proved to be the most suitable for the growth of
bulk crystals of LN, over all the available techniques. The structural defects in LN
crystals have been characterized by various techniques (Wenbo et al., 2004; Wallace
et al., 1970; Sidorov et al., 2007; Sugh et al., 1973; Ivanova et al., 1980; Wicks et al.,
1968). High resolution X-ray diffractometry (HRXRD) being non-destructive in
nature is the most suitable technique to characterize bulk single crystals particularly,
possessing the low and very low angle structural grain boundaries which in general
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 90 ~
arise in the crystals during growth or post growth cooling process
(Bhagavannarayana, Budakoti et al., 2005). As described in earlier chapter (Chapter-
2) LN undergoes the structural changes at its curie temperature i.e. ~1416 K, hence
during the cooling cycle the defects may generate in grown crystals.
During the growth of LN crystals at very high temperature Li gets evaporated
form the melt and hence affect the composition (stoichiometry) and the crystals
generally grow in congruent phase with prominent Li vacancy defects (VLi). When the
same charge is used many times the grown crystals possess slightly yellow color
which may be due to the expected high concentration of VLi defects. Therefore,
special care has to be taken for the growth of bulk single crystals of pure as well as
doped LN. The dopants influence the photorefraction behavior of crystals, therefore,
in parallel to the defect analysis it is essential to study optical parameters
(transmittance, band-gap, refractive indices, birefringence etc). The pure and Zn-
doped LN (Zn:LN) single crystals have been grown by CZ technique. Crystalline
perfection of the grown crystals has been investigated by HRXRD. The crystals have
been extensively characterized by powder X-ray diffraction, Raman, Fourier
transform infrared (FTIR), UV-VIS-NIR, prism coupler and ellipsometry techniques.
To see the effect of repeated use of the same charge, the band gap and in particular the
concentration of VLi and NbLi4+
antisitic defects has been evaluated for such a crystal
which possesses yellow color.
4.2 CRYSTAL GROWTH
The LN and Zn:LN bulk single crystals have been grown by CZ technique.
The details about the crystal puller and growth process are provided in Chapter 3. The
photographs of the harvested crystals are shown in Fig. 4.1(a)-(d). All the obtained
pure as well doped crystals at the time of harvesting from the system were visibly
quite transparent and free from the cracks. The crystal LN and Zn:LN shown in Fig.
4.1(a) and (d) are grown using the fresh LiNbO3 charge which are quite transparent
and colorless, whereas, the LN crystals shown in Fig. 4.1(b) & (c) are grown from the
multiple time used LiNbO3 charge and possess the slight yellow color, the yellow
color may be due to the presence of high concentration Li vacancies.
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 91 ~
10 mm
LN
Fig. 4.1(a): Photograph of CZ grown LiNbO3 single crystal using the seed crystal of [001]
direction, from the fresh LiNbO3 charge
Fig. 4.1(b): Photograph of CZ grown LiNbO3 single crystals using multiple times used LiNbO3
charge, the seed of the same set [Fig. 2.6(b)] has been used
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 92 ~
35 mm
50
mm
31 mm
LN
10 mm
Zn:LN
Fig. 4.1: (c) The Photograph of CZ grown LiNbO3 single crystal from the multiple used charge
and (d) the photograph of Zn:LiNbO3 single crystal grown from the fresh LiNbO3 charge
LN
Zn:LN
Fig. 4.1: (e) the photographs of polished Z-cut wafers (0.5 mm) of pure and Zn-doped (1 mol%)
LiNbO3 single crystals, the letters in background of wafers visualize their high transparency
This is expected because the Li is a light metal and gets evaporated during the crystal
growth process at high melting point. The complete description about the Li vacancies
and antisitic defects in pure and doped crystals has been discussed in the following
sections. For doping of the crystal 1.0 mol% concentration of ZnO was opted which is
well below its threshold concentration (~5.5 mol%). For the characterization purpose
of the structural defects and other physical properties the crystals have been cut to
obtain the Z-cut wafers of the appropriate size from the respective crystal boules. The
wafers were lapped to the thickness of ~500 μm and optically polished. The two such
(c) (d)
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 93 ~
typical wafers are shown in Fig. 4.1(e) and it is clear from the photograph that crystals
are visibly quite transparent and do not possess any physical cracks.
4.3.1 CHARACTERIZATION STUDIES
To study the crystal structure/ crystal system and effect of dopant on the
structure and unit cell parameters the powder X-ray diffraction (PXRD)
measurements on the powdered specimens of LN and Zn:LN crystals were performed.
The detailed description of the diffractometer is given in §3.2. The homogeneous
powdered specimens of the crystals were prepared by crushing the small portions
from the pure as well as doped crystals. The diffraction spectra for both the crystals
were recorded in 20–80 degree angular range of 2-theta, with step size 0.01º and
time/step of 0.1 s.
For evaluation of the crystalline perfection of grown bulk crystals, a
multicrystal X-ray diffractometer (MCD) has been employed the detail description
about of MCD is provide in §3.3. The rocking curves (RCs) have been recorded for
(006) diffraction planes of Z-cut crystal wafers. The specimen was rotated about the
vertical axis, perpendicular to the plane of diffraction, with minimum angular interval
of 0.4 arc sec. The RCs were recorded in ω-scan, in symmetric Bragg geometry. The
X-ray power, size of the beam and diffraction configuration of the diffractometer was
kept constant for both the specimens during all the experiments. The results have been
described in detail in the forthcoming section.
The Raman spectral measurements of the grown crystals have been carried out
on Renisha inVia Raman Microscope in back-scattering mode at the room
temperature. The details about the spectrometer are given in §3.5. A 785 nm
wavelength diode laser (HP-NIR) was used as source with beam power of 300 mW.
The spot size of beam on the crystal surface was 1 µm and the exposure time was kept
10 s. To study the stretching vibrational mode O–H corresponding to the in-diffused
protons (H+) and to evaluate their concentration the Fourier transform infrared
spectrum of the grown crystal in transmission mode has been recorded on a FTIR
spectrometer (Nicolet-5700) [§3.4] in the wavenumber range of 400 – 4000 cm-1
. The
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 94 ~
UV-VIS-NIR transmission and absorption spectra for both the crystals were recorded
using a Shimadzu-1601 spectrophotometer in the wavelength range of 200–1100 nm.
The optical transparency, band gap, and the composition of crystals have been
evaluated.
Refractive indices (ordinary and extraordinary) and birefringence of the Z-cut
samples were measured on a Prism coupler spectrometer (PCS) for 1064 and 532 nm
wavelengths. The wavelength dispersion of both the crystals has been studied by
ellipsometry [§3.9]. The average refractive index and extinction coefficients for the
wavelength range of 200 – 1000 nm on have been evaluated.
4.3 RESULTS AND DISCUSSION
4.4.1 Powder X-ray diffraction analysis
The LiNbO3 crystals have the rhombhohedral structure at room temperature
and belong to the R3c space group 3m point group. In this space group lattice of
LiNbO3 crystal Nb atoms occupy (0,0,0) positions, Li atoms occupy (0,0,1/4)
positions and O atoms occupy (4,1/3,1/2) positions. The lattice parameters of the
hexagonal unit cell of LiNbO3 are a = 5.1483 Å and c = 13.8631 (Abrahams, Reddy et
al., 1966) Å. LiNbO3 undergo the first order phase transition at ~1160°C, due to
which it transforms from the polar ferroelectric 3m symmetry to nonpolar paraelectric
3m symmetry, above the Curie temperature. At this phase transition a substantial
change in the volume of the unit cell take place (Parlinski et al., 2001) due to which
strains are produced in the crystal lattice. Therefore at this phase transition
temperature the growth conditions are to be handled very carefully during their
cooling cycle after the growth, otherwise the crystals undergo the severe problem of
cracking when they are taken out of the furnace. Li is a light element and during the
growth process of crystals it undergoes the evaporation and hence the grown crystals
are deficient in Li. This results in, the missing of Li+ ions from their regular sites and
most of these Li vacant sites in the lattice are occupied by Nb5+
and result in the
NbLi4+
point defects, termed as antisitic defects. For the confirmation of crystal system
and evaluation of strains in the lattice of crystals the PXRD spectra of for the
powdered specimens of LN and Zn:LN crystals were recorded and shown in Fig. 4.2.
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 95 ~
0
800
1600
2400
3200
20 30 40 50 60 70 800
800
1600
2400
(312)
(306)
(220)
(208)
(202)
(110)
(214
)
(300)
(018)(1
16)
(113)
(024)
(122
)
(104)
(006)
LN
(012
)
Inte
nsi
ty (
a.u
.)
2-theta (degree)
Zn:LN
Fig. 4.2: The indexed PXRD spectra of pure and Zn-doped LiNbO3 crystals
0.0005
0.0010
0.0015
0.0020
0.2 0.3 0.4 0.5 0.60.0005
0.0010
0.0015
0.0020
LN
Linear Fit
c
os
= -7.54x10-4
= 9.38x10-6 Zn:LN
Linear Fit
c
os
sin
Fig. 4.3: β cosθ vs. sinθ plots for for an evaluation of lattice strain (η) in the lattice of crystals
The diffraction peaks are very sharp which indicate good crystallinity of the grown
crystals. All the peaks occurred in LN spectra are also present in Zn:LN, and reveal
that the grown crystals possess the perfect single phase with trigonal (rhombohedral)
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 96 ~
crystal structure and R3c (C3ν6) space group (Abrahams et al., 1966; Chemaya et al.,
2001).
The strain (η) in the crystal lattice have been evaluated by using Hall-
Williamson relation [§3.2]. The βcosθ vs. sinθ plots are shown in Fig. 4.3. The
evaluated values of η for LN and Zn:LN are respectively -7.54×10-4
and 9.38×10-6
.
The higher –ve value of η for pure LN crystal is owing to the presence of vacancy
type of defects (VLi) in LN which cause expansion of lattice around the defect core
and lead to the tensile strain (Bhagavannarayana, 1989). For Zn:LN the simulated plot
is almost horizontal having very low +ve value of η which infers that the Li vacancies
(VLi) have been removed by doped Zn2+
ions at low concentration (~1 mol%) and due
to their large size over Li+ lead to the slight compressive strain in the vicinity of
defect core. Such low strain values depict that doped crystal is almost free from the
strains and the strains of such order in nearly perfect silicon single crystal was
evaluated by diffuse X-ray scattering technique at authors’ laboratory
(Bhagavannarayana 1989)29. The antisitic NbLi4+
defects in LN have been reported to
be completely wiped out for the threshold concentration (~5.5 mol%) of Zn doping
(Chemaya et al., 2001). The detailed crystalline perfection analysis of both the grown
crystals is given in the flowing §4.4.2 of HRXRD.
4.4.2 High resolution X-ray diffraction analysis
For better understanding about the crystalline perfectness of the pure and
doped single crystals of LiNbO3 in comparison with the ideal (theoretical) LiNbO3
single crystal, the theoretical rocking curves for (006) diffraction planes have been
obtained and shown in Fig 4.4. These diffraction curves have been obtained by
considering the plane wave theory of dynamical X-ray diffraction (Batterman & Cole,
1964) for an ideally perfect crystal. The figure contains two diffraction curves; one
the so called Darwin, where the phenomenon of linear absorption of X-rays is not
taken in to consideration and the other, the well-known Darwin-Prince curve in which
the linear absorption is taken into account. The HRXRD RCs for (006) diffraction
planes of the Z-cut wafers of grown LN and Zn:LN bulk crystals are recorded using
ω-scan in symmetrical Bragg diffraction geometry and are shown in Fig. 4.5.
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 97 ~
-20 -10 0 10 200.0
0.5
1.0
Perfect
=1.38x10-5 rad
PFH
B
2sin
22
1
LiNbO3
(006) planes
MoK1
2.6"
Darwin
Darwin-Prince
Ref
lect
ivit
y, I/
I 0
Glancing angle , [arc sec]
Fig. 4.4: The Darwin and Darwin-Prince theoretical rocking curves for (006) diffraction planes
of pure LiNbO3, generated using the plane wave theory of dynamical X-ray diffraction
The RC of LN is having a single peak with FWHM value of 62 arc sec, which is quite
higher than that of the theoretical value as shown in Fig. 4.4 (Kushwaha, et al., 2011).
The higher value of FWHM of recorded RC for LN indicates the presence of the point
defects, which are expected due to expected deviation from the stoichiometry of real
life LN crystals. The absence of any additional peak in RC depicts that the grown LN
is almost perfect and free from the structural grain boundaries which are very
common to observe in LN crystals (Bhagavannarayana et al., 2009).
However, asymmetry of the diffraction peak with high scattered intensity
along -ve side in comparison to that along +ve side with respect to the exact Bragg
peak position indicates that the LN crystal predominantly contains vacancy type of
defects (Kushwaha et al., 2011). This is due to the fact that around the vacancy
defects the lattice undergoes tensile stress and the lattice parameter d (interplanar
spacing) increases which results in high scattered intensity (also known as diffuse X-
ray scattering) at slightly lower (Δθ) glancing angles with respect to the exact Bragg
angle (θB). This is because d and sinθB are inversely proportional to each other in the
Bragg equation (2dsinθB = nλ; n and λ being the order of reflection and wavelength of
X-rays respectively, which are fixed). It is important to mention here that the variation
in the lattice parameter (d) is confined only in the vicinity of point defect which gives
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 98 ~
the scattering intensity only close to the Bragg peak position. Li is a light element and
during the growth process at high temperature (1523 K) it evaporates and lead to
change in the stoichiometry of the grown crystals with lattice defects VLi and NbLi4+
(Nb5+
ions at Li sites) antisitic defects (Chemaya et al., 2001; Nagashio et al., 2004).
However, the expected change in the stoichiometry is very meager, and LN crystal is
congruent in nature with a small deficiency of Li. This result is supported by the
calculation of Li concentration using Raman and birefringence analysis.
The RC for Zn:LN [Fig. 4.5] has quite different features with scattered
intensity in a broad angular region (~300 arcsec) of glancing angle. Instead of a single
peak as for LN, it consists a shoulder (as an inseparable peak merged with the main
peak) at the +ve side of the peak position. This indicates that the grown Zn:LN crystal
possesses a low angle grain boundary. On de-convolution of the curve we get two
peaks separated by 70 arcsec. The solid curve is the convoluted curve obtained by the
Lorentzian fit of the data points and it is well fitted with the experimental points. The
main peak with FWHM of 86 arcsec corresponds to the main crystal domain and the
smaller peak with FWHM 70 arcsec is due to the low angle internal structural grain
boundary which is mis-oriented from the main crystal domain by 70 arcsec. Such
structural grain boundaries are very common to observe in real crystals particularly in
ferroelectric crystals, which undergo structural phase transitions associated with
volume changes, during post-growth cooling cycle.
The dopants in crystal generate heavy stresses in the lattice and the lattice tries
to relax which results in the formation of dislocations and boundaries. Segregation of
dopants takes place along the grain boundaries through the process of guttering
(Bhagavannarayana & Kushwaha, 2010; Saraev et al., 2003; Grange et al., 1990). The
radius of dopant ions Zn2+
is In the present case of Zn doped crystal Zn:LN, when the
Zn2+
ions (74 pm) comparable to Li+ ion (76 pm) and larger than that of Nb
5+ (64 pm)
occupy the VLi or Nb 4
Li positions in the crystal, the lattice around the defect core
undergoes compressive stress as observed in §4.4.1 of PXRD. Such type of structural
grain boundaries were successfully investigated by HRXRD and X-ray topography in
pure LN which could be removed successfully by the post growth annealing process
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 99 ~
0
100
200
300
400
-300 -200 -100 0 100 200 3000
100
200
300
400
Vacancy
defect
62"
LN (006) planes
MoK1
(+,-,-,+)D
iffr
acte
d X
-ray I
nte
nsi
ty [
c/s
]
Grain-
Boundary
70"
70"Zn:LN (006) planes
MoK1
(+,-,-,+)
86"
Dif
fracte
d X
-ray i
nte
nsi
ty [
c/s
]
Glancing angle [arc sec]
Fig. 4.5: The RC’s recorded for the (006) diffracting planes of LN, the inset indicates the vacancy
defect (VLi) with expansion of the lattice around the defect core, and of Zn:LN, the inset shows the
schematic of a low angle grain boundary (α << 1’)
with very low rates of heating and cooling (Bhagavannarayana et al., 2009). The
HRXRD analysis indicates that both the grown crystals (LN & Zn:LN) are almost
perfect and decipher their suitability for device applications. Though the specimens
contain a low angle boundary, the relatively low angular spread of around 300 arcsec
of the diffraction curve and the low FWHM values show that the crystalline
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 100 ~
perfection is reasonably good. It may be mentioned here that such low angle
boundaries (which are unlikely to degrade the properties) could be detected with
resolved peaks in the diffraction curves only because of the high resolution of the
diffractometer, characterized by very low values of wavelength spread (i.e. ∆λ/λ ≈
10-5
) and horizontal divergence (<< 3 arcsec) for the exploring or incident beam.
4.4.3 Raman spectroscopy analysis
The LiNbO3 primitive unit cell with two formula weights (10 atoms) results in
30 degrees of freedom, 27 of which assigned as phonon modes and other three as
acoustic phonons. The optical modes of LN for R3c are given by the relation:
EAAoptical 954 21 . Only A1 and E modes are Raman active therefore only 13
phonon peaks are expected in the spectra, the degeneracy between longitudinal optical
(LO) and transverse optical (TO) phonons has been lifted due to the long-range
electrostatic fields because of ionic nature of LN (Scott et al., 2004). The Raman
spectra recorded for the pure as well as doped Z-cut single crystal wafers are shown in
Fig. 4.6 and the vibrational frequencies with assignments are given in Table 4.1. Both
the spectra consist of similar features with the all fundamental vibrational modes
corresponding to LiNbO3 structure. The peak parameters of the Raman spectra are
known to be very sensitive for the structural changes of LiNbO3 crystals, particularly
those induced by the deviations from stoichiometry and structural defects (Sidorov et
al., 2007). Significant differences in the peak intensities of the recorded spectra for
both the crystals have been observed, with no significance change in the peak
positions. The intensity of the line at ~186 cm-1
is said to be strongly dependent on
structural defects and its low intensity for LN may be attributed to the presence of
stresses in the lattice owing to the VLi defects revealed by PXRD and HRXRD. The
relaxation of the lattice from strains of Zn:LN crystal might have led to the increase in
the intensity of this peak. The increase in the intensity of the other peaks also may be
attributed to the same reason.
The observed modes in the present crystals are in well agreement with various
reports available in the literature (Scott et al., 2004; Sidorov & Palatnikov et al.,
2003; Schlarb et al., 1993). Except the variation in the intensities of the peaks, there is
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 101 ~
200 400 600 800 10000.0
30.0k
60.0k
90.0k
11.28 cm-1
LN
Zn:LN
150 cm-1
12.03 cm-1
878 cm-1
28.83 cm-1
28.84 cm-1
878
Inte
nsi
ty (
a.u.)
Raman shift (cm-1)
LN
Zn:LN
150
Fig. 4.6: The Raman spectra for both the crystals recorded in back scattering mode shows all
thirteen vibrational modes. At top shows the magnified view of peaks at 150 and 878 cm-1
the line-
widths (Γ) of these used to evaluate the composition of the crystals
no inconsistency in the spectra due to Zn doping and reveals no change in the basic
structure of crystal lattice as analyzed by PXRD and HRXRD. The E(TO) and A1(LO)
modes respectively at ~150 & 878 cm-1
are used to calculate the Li concentration
(CLi) in both the grown crystals using the empirical formulae: CLi[mol%] = 53.03 -
0.4739Г [cm-1
] for 150 cm-1
and CLi[mol%] = 53.29 - 0.1837Г [cm-1
] for 878 cm-1
(Schlarb et al., 1993) indicated by arrows in figure. Here Г is the linewidth (full width
at half maximum) of the peak, its values for the mode E(TO) for LN and Zn:LN are
respectively 11.28 and 12.03 cm-1
, whereas for A1(LO), these are 28.84 and 28.83
cm-1
. The slight variation in Г is owing to the change in translational symmetry due to
the presence of point defects (VLi, NbLi and ZnLi), which influence the interionic
potential and hence the selection rules of transitions (Chemaya et al., 2001; Schlarb et
al., 1993). Evaluated CLi values are given in Table. 4.1, and indicate that the grown
crystals have the congruent composition
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 102 ~
Table 4.1 Raman scattering modes and evaluated CLi for LN and Zn:LN bulk single crystals
(Li/Nb = 0.954 & 0.948 respectively for LN and Zn:LN), slightly off from the
stoichiometric composition (Li/Nb = 1) (Scott & Burns, 1972; Balanevskaya et al.,
1983). The CLi results for LN with deficiency of Li indicating the congruent nature of
crystal is in correlation with PXRD and HRXRD analysis in §4.4.1 & §4.4.2
indicating the presence of tensile stresses and vacancy defects.
4.4.4 Fourier transform infrared analysis
In LiNbO3 crystals, due to thermal diffusion of H+ ions, the fixed holograms
have a non-infinite lifetime at room temperature. However, this life time is too longer
to make room temperature direct measurements and hologram lifetime is found to be
a function of the H+ ion concentration (H0), of the concentration of iron and relative
valance ratio, and of the diffusion coefficient of H+ ions (DH) through the
relation,
1
0
2
12
t
FN
H
, where Ʌ hologram fringe spacing. So, low
hydrogen concentration and large Nt (=[Fe2+
][Fe3+
]/[Fe]) are required for long lasting
holograms (Arizmendi et al., 2004; Miguel et al., 2002).
Modes of vibration ν (cm
-1) CLi (mol%)
LN Zn:LN LN Zn:LN
E(TO) 150 149 47.7 47.4
A1(TO) 186 187 - -
E(TO) 235 235 - -
A1(TO) - 253 - -
E(TO) 260 - - -
A1(LO) 271 270 - -
E(LO) - 296 - -
E(TO) 321 321 - -
E(TO) 367 367 - -
E(TO) 431 431 - -
E(TO) 579 578 - -
E(TO) 629 628 - -
A1(LO) 878 879 48.0 48.0
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 103 ~
4000 3000 2000 10000
20
40
60
80 LN
Zn:LN
Wavenumber (cm-1
)
Tra
nsm
itta
nce
(%
)
3520 3480 3440 3400
33 cm-1
31 cm-1
Fig. 4.7: The FTIR spectra of pure and Zn-doped recorded in transmission mode and the inset in
figure shows magnified view of absorption band corresponding to O-H stretching vibration.
Therefore, in parallel to the defect analysis of the grown crystals, it is very important
to analyze them for the presence of OH- ions. The recorded FTIR spectra of pure as
well as doped crystals are shown in Fig. 4.7. In both spectra the broad bands at ~3484
cm-1
indicate the presence of OH- ions related defects (Klauer & Wӧhlecke, 1994) and
the FWHM of the peaks for LN and Zn:LN are respectively 33 and 31 cm-1
. The
magnified view of these bands is shown in the inset of figure. Although the absorption
bands do not have any separated peaks but have comprised of a clear absorption
hump, at approximately ~14 cm-1
away from the main peak position of ~3468 cm-1
.
The relative intensities of the observed peaks are reported to be independent of the
temperature (Herrington et al., 1973).
The spectra infer the aspects about the incorporation of H+ into the crystal
lattice that the transitions are of electric dipole nature and the band is the 0–1
stretching vibrational band of OH- and there are two different transitions near 3484
cm-1
(2.87 μm). It is also reported that the energy levels are not strongly coupled to
phonons (Herrington et al., 1973). Due to relatively large polarizabilities of the O2-
ions and the coulomb forces occupying an anion site, the OH- bond is believed
parallel to the shortest O2-
–O2-
bond favored. The OH bond stretching, normally in
inorganic and organic crystals, is with the energies of the order of 3000–3100 cm-1
.
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 104 ~
Oxygen
Nb
proton
Li
below plane above plane
1
23
6
45
3540 3510 3480 3450 34200.00
0.25
0.50
LN
Zn:LN
(cm-1)
(
cm-1)
Fig. 4.8: The schematic of an oxygen (001) plane in hexagonal unit cell indicating different O–O
bond lengths and possible lattice cites feasible for proton incorporation (Klauer & Wӧhlecke,
1994). Nature of the absorption constant for OH- absorption band occurred in FTIR spectra
But in LiNbO3 the energy of the band is comparatively higher and only slightly less
than that of free hydroxyl ions (3650 cm-1
), which indicate that OH- ions form very
weak hydrogen bonds only. Figure 4.8 shows the schematic for the nearest neighbors
of O2-
anion in LN. As it is clear from the schematic the oxygen atoms are in the
distorted closed-packed hexagonal structure of LN due to which the sites of hydroxyl
ions are nonequivalent and hence result in components of OH band in spectra. This
stretching energy difference is reported to be very less (~13 cm-1
) (Herrington et al.,
1973).
In our crystals as the peaks in the absorption band are not resolved and form
only a hump which indicates that the crystals’ lattice is not distorted to a great extent
and therefore the transitions corresponding to the O-H stretching of the incorporated
protons at different sites in the lattice are not distinguishable, the hump in case of the
Zn:LN is seem to be diminished and the FWHM of the entire band reduced to 31 cm-1
than that of 33 cm-1
for LN crystal. Such feature of the absorption band indicates that
the distortion of the crystal lattice decreased with Zn doping, by diminishing the
vacancies due to Li ion vacancies and Nb 4
Li antisitic defects. This is well understood
by the evaluation of OH- ions concentration incorporated in the lattice by using the
absorption constant ‘κ’ corresponding to the maxima of bands (shown in Fig. 4.8).
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 105 ~
The concentration of OH- ions was evaluated by using the relation: n(OH
-) = (3 ± 1) ×
1019
κ(3484 cm-1
) cm-3
, here n is the density of the OH- ions and these have been
found to be 1.17×1019
ion/cm3 and 0.97×10
19 ion/cm
3 respectively for LN and Zn:LN
(Bollmann et al., 1976; Bollmann & Stohr, 1977; Herrington et al., 1973). The lower
OH- ion concentration in Zn:LN indicates that due to the occupancy of VLi and
Nb 4
Li sites by Zn2+
ions, the diffusion of H+ into the lattice of Zn:LN reduced. This
result is in good agreement with the evaluated strains in the crystals’ lattice by PXRD
and Raman investigations. The band at 1746 cm-1
is owing to the characteristic Nb–O
overtone band. Such features of the FTIR spectra assure the perfect molecular
structure of grown crystals and display a good correlation with crystalline perfection
assessed by HRXRD and PXRD techniques and will be further verified by UV-VIS
studies as follows.
4.4.5 UV-VIS-NIR optical analysis
The optical transmission spectra are recorded in the wavelength range of 250-
1100 nm and are shown in Fig. 4.9. The spectra depict that both the crystals are highly
transparent in the entire visible region. In comparison to the pure crystal Zn-doped
crystal possesses slightly higher transparency in the entire visible as well as UV
regions. This enhancement in transparency due to Zn doping is attributed to increase
in the photoconductivity (Simon et al., 1995). The calculated absorption coefficient
(α) exhibits the interesting feature with blue shift of the cut-off wavelength to 324 nm
for Zn:LN compared to that of 321 for LN. This shows increase in optical band gap of
LN due to Zn doping. This behaviour of Zn-doping is just opposite to that of Fe
doping as we have observed in our recent studies on Fe:LN, which exhibits red-shift
(Kushwaha et al., 2011). This shift in the absorption edge may be elucidated
by rZ /2* , here sZZ *; Z*, Z, s , and r are the effective nuclear charge
number, the atomic ordinal number of the ion, the shield factor, and the radius of ion,
respectively. The valance electron transition energy from 2p orbits of O2-
to 4d orbits
of Nb5+
decides the absorption edge in LiNbO3 (Zhen et al., 2003), therefore the
position of absorption edge considerably get affected by the symmetry of the valence
electron shells.
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 106 ~
250 375 500 625 750 875 10000
25
50
75
100
300 310 320 330 340 3500
25
50
75
100
3.7 3.8 3.9 4.0 4.1 4.20
6
12
18
T (
%)
(nm)
LN
Zn:LN
(
cm-1
)
(nm)
LN
Zn:LN
(h)2
x1
08 (
eVm
-1)2
h (eV)
LN
Zn:LN
Fig. 4.9: From top to bottom; the optical transmittance (T) spectra in the entire UV-VIS-NIR
wavelength range, absorption coefficient (α) vs. wavelength (λ) plots, λ20 (wavelength
corresponding to 20 value of α) used to evaluate the crystals’ composition and defectiveness, and
(αhν)2 vs. hν plots used for optical band gap ‘Eg’ evaluation
The dopant ions with higher ability to polarize O2-
ion, relative to the original lattice-
site ion, decreases the transition energy of valance electron and vice-versa.
In the present case, the Zn2+
ions with lower Z* and hence lesser ability to
polarize the O2-
ions in comparison to that of Nb5+
, the replacement of Nb 4
Li (present
in pure crystal) by Zn2+
results in decrease the polarization ability of O2-
ions
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 107 ~
300 350 400 4500
5
10
15
20
25
(cm
-1)
(nm)
LN - Yellow
20
= 336 nm
3.0 3.2 3.4 3.6 3.8 4.00
2
4
6
8
10
(
h)2
x10
7 (
eVm
-1)2
LN - Yellow
h (eV)
Fig. 4.10: The absorption coefficient, cut off wavelength and (αhν)2 vs. hν for the yellow crystal,
λ20 has been used to evaluate the defectiveness of crystal
in the consequence of it the energy required for transition of valance electron
increased (Zhen et al., 2003) and hence increases the band gap of crystal. It is
important to mention here that such behaviour of Zn doping is true for its
concentrations below the threshold value. The Fig. 4.10 shows absorption coefficient
spectrum of the grown yellow crystal [Fig. 4.1(c)]. The absorption edge for this
yellow crystal is significantly shifted towards the longer wavelengths. This behavior
of the absorption edge may be due the presence of VLi and NbLi4+
with high
concentration. The direct band gaps of LN and Zn:LN crystals have been calculated
using the Tauc’s relation:
)()( 2 hEAh g (4.1)
and by plotting (αhν)2 vs. hν (Fig. 4.10), where α is the absorption coefficient, ν is the
frequency of incident radiation, h is the Plank’s constant, Eg represents the optical
band gap and A is a constant (Kushwaha, Maurya et al., 2011). The Eg values have
been evaluated by extrapolating the linear part of the plots to the abscissa (hv) as
shown in figure and found to be 3.91 and 3.97 eV respectively for LN and Zn:LN.
This increase in the band gap of LN by Zn-doping at 1 mol% (which is well
below the threshold concentration of ~6 mol%) could be attributed to the occupancy
of Zn2+
ions at Li+ sites in the form of Zn which are otherwise occupied by Nd
5+ in
the form of Nb 4
Li (Zhen, Xu et al., 2002). At high concentrations, Zn facilitates LN
for excellent photorefraction applications in UV region of electromagnetic radiation
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 108 ~
Table 4.2 Evaluated Li/Nb, C(VLi) and C(NbLi) values for the grown crystals
(Haijun et al., 2004). From the Fig. 4.10 it is clear that the band gap for the yellow
crystal is very less than that of transparent pure LN crystal.
The Li/Nb ratio for the grown crystals has been evaluated by using the
empirical formula,
2
20
29.81
54.3011
Nb
Lisuitable for nominally pure as well
doped crystals with concentration below threshold value, here λ20 is the wavelength
(nm) corresponding to the absorption coefficient of 20 cm-1
. The Li/Nb values are
found to be 0.954 and 0.959 respectively for LN and Zn:LN crystals and used to
estimate the concentration of VLi and NbLi which act as defects in crystals, using the
formulae (4.2) for pure and (4.3) & (4.4) for doped crystals (Salloum et al., 2010).
%)(100)/(5
)/(44)( mol
NbLi
NbLiVC Li
& 4/)()( LiLi VCNbC (4.2)
%)(100)/(5
)/)(4(34)( mol
NbLi
NbLiyyVC Li
(4.3)
%)(100)/(5
)/(21)( mol
NbLi
NbLiyNbC Li
(4.4)
where, y is the molar percentage of dopant in the crystal, which is 1 mol% ZnO in the
present case. The evaluated defect concentrations are given in Table 4.2. The lower
concentrations of VLi and NbLi in doped crystal indicate that the doped crystal
comprises lesser point defects compared to that of pure crystal, as illustrated by
PXRD and HRXRD, lead to the fall in the lattice strains. The reduced concentration
of NbLi manifests the slight increase in the band gap. The significant blue shift of the
absorption edge in stoichiometric as well congruent crystals and interesting
photorefractive behavior in UV region has been reported with varied Mg doping
(Salloum et al., 2010; Jingjun et al., 2000). The value λ20 for yellow crystal is very
Crystal λ20 (nm) Li/Nb C(VLi) (at %) C(NbLi) (at %) Eg (eV)
LN 319 0.954 0.62 0.15 3.91
Zn:LN 318 0.959 0.48 0.07 3.97
LN-Yellow 336 0.820 2.47 0.62 3.56
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 109 ~
high compared to that of LN as well as Zn:LN. The ratio of Li/Nb for yellow crystal is
found to be significantly less i.e. 0.820. The tabulated values of C(VLi) and C(NbLi)
[Table 4.2] are very high compared to those of LN and Zn:LN and these results
clearly decipher that the grown yellow LN crystal contains more Li vacancies and
NbLi antisitic defects and hence these results clearly reveal that when we use the same
lithium niobate residual charge many times for the growth of single crystals, the
grown crystals have abundance of intrinsic defects and get yellow color due to color
centers at Li vacancies.
4.4.6 Prism coupler birefringence analysis
Lithium niobate is an excellent nonlinear optical (second harmonic generation:
SHG) crystal due to its birefringence nature. The Nb 4
Li and VLi intrinsic defects in LN
influence the photorefraction behavior by manipulating photo-induced charge
transport and thereby the change in refractive index. The NbLi act as electron shallow
traps and VLi as hole traps. The birefringence of lithium niobate crystals increases
significantly in the presence of transition metals as dopants (Buse et al., 1998;
Peithmann et al., 2002; Peithmann et al., 1999; Haijun et al., 2004; Koster et al.,
2009; Jaque et al., 2004). The ordinary (no) and extraordinary (ne) refractive indices
of grown LN and Zn:LN crystals have been measured for the wavelengths 532 and
1064 nm and are plotted in Fig. 4.11 and the actual values are tabulated in Table 4.3.
The measured refractive indices are in well agreement with earlier reports (Schlarb et
al., 1995; Weis et al., 1985; Volk et al., 2001). The refractive index values are found
to be higher for 532 nm compared to those for 1064 nm. The no for both pure and Zn-
doped LN crystals are almost same at both the wavelengths 532 nm and 1064 nm,
whereas ne changed in doped crystal to lower values. The birefringence
( oe nnn ) of crystals are found to be higher for LN in magnitude at 532 nm in
comparison at 1064 nm. In contrast to LN, ∆n for Zn:LN found to be increased at 532
nm whereas at 1064 nm, it is decreased. This may be due to the fact that the higher
energy photons are able to excite the charge carriers from deeper level rather than the
photon of lower energy and lead to the photorefraction in LiNbO3 in ultraviolet
wavelength range (Haijun et al., 2004). The ∆n values at different wavelengths of the
pure crystal for have been used to evaluate the stoichiometry (CLi) of LN crystals
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 110 ~
2.226
2.268
2.310
2.352
no
LN
Zn:LN
2.142
2.184
2.226
2.268
ne
532 1064-0.056
-0.070
-0.084
-0.098
n
Wavelength (nm)
Fig. 4.11: The ordinary (no), extraordinary (ne) and birefringence (∆n) parameters of pure LN
and Zn:LN crystals at 532 and 1064 nm wavelengths measured by the prism coupler spectrometer
Table 4.3 The evaluated no, ne, ∆n and CLi values for the grown single crystals
Crystal λ (nm) no ne Δn CLi (mol%)
LN 532 2.324 2.234 -0.090 48.13
1064 2.231 2.156 -0.075 47.92
Zn:LN 532 2.323 2.227 -0.096 -
1064 2.233 2.167 -0.066 -
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 111 ~
using the relation nbaCLi )()( , where a(λ) and b(λ) are the wavelength
dependent parameters and the values of these parameters are taken from the literature
for the respective wavelength used to measure birefringence and CLi values are listed
in Table 4.3. The CLi for LN are of the same order as obtained by the Raman analysis
and indicate that the grown crystals are only slightly off from their stoichiometry and
is congruent in nature. The PXRD and HRXRD structural analysis along with their
optical behaviour infer the congruent nature of grown crystals which is further
evidenced by the composition evaluation.
4.4.7 Wavelength dispersion analysis
The wavelength dispersion analysis of the grown crystals is analysed by
ellipsometry and the optical constants have been evaluated. The optical constants
define how light interacts with a material. The complex refractive index is a
representation of the optical constants of material and is represented by ñ = n + iκ.
The real part ‘n’ is the index of refraction, defines the phase velocity of light in
material: v = c/n, where v is the speed of light in material and c is the speed of light in
vacuum. The imaginary part ‘κ’ is the extinction coefficient, determines how fast the
amplitude of the wave decreases. The extinction coefficient is directly related to the
absorption of material and is related to the absorption coefficient by: α = 4πκ/λ,
where, α is the absorption coefficient and λ is the wavelength of light. The measured n
and κ are plotted as a function of wavelength for LN and Zn:LN crystals and are
shown in Fig. 4.12. The obtained average RI values for both the crystals are of the
order of reported in literature and in the above section evaluated by prism coupler
measurements. The RI values are higher at lower wavelengths compared to that in
high wavelength region. In the high wavelength region RI values for both the crystals
are similar, whereas at lower wavelengths these values are higher for Zn:LN crystal.
This behaviour of RI for doped crystal is, because Zn enhances the photorefraction
behaviour of LN in the ultraviolet region by removing the localized energy states
present in the optical band gap caused by the vacancies and antisitic defects in the
lattice. The κ increases with increase in the energy and above 300 nm it remains
almost constant, it is clear from the plots that κ for Zn:LN is slightly lesser over the
entire energy range.
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 112 ~
2.25
2.50
2.75
0.30 0.45 0.60 0.75 0.900.0
0.5
1.0
n
LN
Zn:LN
(m)
LN
Zn:LN
Fig. 4.12: The variation of average refractive index ‘n’ and extinction coefficients ‘κ’ with
wavelength recorded for (006) planes of grown pure and Zn-doped LiNbO3 single crystals
4.5
5.4
6.3
7.2
0.30 0.45 0.60 0.75 0.900
2
4
r
LN
Zn:LN
i
(m)
LN
Zn:LN
Fig. 4.13: The variation of real ‘εr’ and imaginary ‘εi’ components of the dielectric constant
with the wavelength. These components have been evaluated from n and κ parameters
The linear response of the matter to electromagnetic radiation is best described
by the complex frequency-dependent dielectric constants (ε = εr + εi ) where, εr and εi
are respectively the real and imaginary parts of optical dielectric constant.
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 113 ~
0 2 4 6 8 10
0.22
0.24
0.26
0.28
(Eo & E
d)
LN= 8.96 & 32.6 eV
LN
Zn:LN
Linear fit - LN
Linear fit - Zn:LN
1/(
n2-1
)
(h)2 (e
2V
2)
(Eo & E
d)
Zn:LN= 7.71 & 26.1 eV
Fig. 4.14: The (n2-1)
-1 vs. (hν)
2 plots for and the linear fit to these plots for the evaluation
of dispersion energy ‘Ed’ and average single oscillator energy ‘Eo’ of LN and Zn:LN crystals
These are related to the more directly measureable parameters, refractive index and
extinction coefficient through the following relations (Wolton et al., 1963; Wemple et
al., 1977):
22 nr and ni 2 (4.5)
The variations of εr and εi with wavelength are shown Fig. 4.13. Both εr and εi have
higher values at lower wavelengths and decrease very fast with increase in
wavelength but for the wavelengths above 400 nm these remain almost constant, and
follow the similar behaviour respectively of n and κ. The εr values for doped specimen
are higher over the lower wavelength range up to 600 nm. However εi values are
almost same for both the crystals. The dispersion plays an important role in the
research for optical materials due a significant factor in optical communication and in
designing the devices for spectral dispersion. The single-oscillator parameters for the
crystals were calculated and analyzed using Wemple-DiDomenico model
(DiDomenico et al., 1969):
22
0
02
)(1
hvE
EEn d
(4.6)
where, h is Plank’s constant, Eo is the average single oscillator energy for electronic
transitions and Ed is the dispersion energy or oscillator strength which measures the
average strength of interband optical transitions. These parameters for both the
crystals have been obtained by plotting 1/(n2-1) vs. (hv)
2 in Fig. 4.14 for the 1 – 10
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 114 ~
e2V
2 range of (hv)
2. The light and dark straight lines are the linear fit to the plots
respectively for pure and doped crystals, have been extended to the ordinate at zero of
abscissa. The linear fit plot for Zn:LN exhibits higher slope and intercept on the
ordinate compared to that of pure. The Eo and Ed have been calculated from the slope
(EoEd)-1
of linear fit curves and intercept the (Eo/Ed) on ordinate cut by the extension
of linear fit. The oscillator strength and dispersion energies of crystals strongly
depend on their structures (Wemple & DiDomenico, 1969). The evaluated Eo and Ed
parameters for pure crystal are slightly higher than that of reported values, however
for the Zn-doped crystal these parameters are in good agreement with reported one.
(Wemple & DiDomenico, 1971). The deviation in these parameters for the pure
crystals can be correlated with the presence of abundance of Li+ ion vacancies and
NbLi4+
antisitic defects in pure crystals as described in §4.4.1 and §4.4.2. However, in
Zn-doped crystal these point defects were removed and resulted in widening of optical
band-gap of crystal and dispersion parameters have attained the perfect values. The
normalized dispersion energy ‘β’ evaluated by a general empirical expression; Ed = β
NcZaNe, where Ne is the effective valance electrons per anion, Za anion valancy and Nc
the coordination number of the nearest cation. Its value for Zn:LN crystal is 0.27 and
falls well in the category ionic oxide crystals, whereas for LN it is 0.34 slightly higher
due to the higher dispersion nature of crystal.
4.4 CONCLUSION
The bulk single crystals of pure and Zn-doped (1 mol%) LiNbO3 have been
successfully grown by indigenously developed CZ puller assisted with radio
frequency heating furnace and a post growth resistive heater. Both the crystals are
visibly quite transparent. The PXRD analysis confirmed the crystal structure and
space group which are same for both the crystals. The powder XRD results revealed
stresses in the crystals and found to be tensile strain in LN due to vacancy defects
whereas compressive strain Zn:LN crystal due to Zn-doping. The HRXRD revealed
that that the crystalline perfection of both pure and doped crystals is reasonably good.
LN contains the vacancy defects owing to the Li deficiency and free from the
macrodefects like grain boundaries, however Zn:LN contains low angle grain
Chapter – 4 LiNbO3 & Zn:LiNbO3
~ 115 ~
boundaries. The Raman analysis for the defects analysis found to be consistent with
PXRD and HRXRD results and illustrate the congruent composition of the grown
crystals. The O–H stretching vibrations in crystals originated due to the incorporation
of protons (H+) in the crystal lattice and have been assessed by FTIR. The
concentration of OH- ions in both the crystals has been evaluated, which is slightly
higher for the pure crystal.
Compared to that of LN the optical transparency of Zn:LN increased over the
entire recorded UV-VIS-NIR spectrum. Band gap was found to be slightly increased
due to Zn doping. The enhanced optical transparency and band-gap envisage the
suitability of Zn:LN for SHG applications. The birefringence of the doped crystals
found to be increased for lower wavelengths however, at higher wavelengths it is
found to be decreased, such behaviour of doped crystal makes it suitable for the
photorefractive holographic data storage applications in the ultraviolet region. The
wavelength dispersion behavior of grown crystals has been revealed that the average
single oscillator energy for electronic transitions (Eo) and dispersion energy or
oscillator strength (Ed) of optcal transitions. The Eo and Ed values found to be strongly
dependent on the crystalline perfection/ lattice defects.
~ 116 ~