Content uploaded by Ji Wang
Author content
All content in this area was uploaded by Ji Wang
Content may be subject to copyright.
Acousto-optic modulation using a new chlorotellurite glass
I. Abdulhalim, C. N. Pannell, J. Wang, G. Wylangowski, and D. N. Payne
Optoelectronics Research Centre, OpticaI Eibre Group, Southampton University, Southampton SO9 SNH,
United Kingdom
(Received 16 July 1993; accepted for publication 21 September 1993)
Results of initial tests on an acousto-optic modulator using a new chlorotellurite glass with a
composition of 3TeO,-2ZnC12 are reported. This glass exhibits a wider transparency range
(0.42-5.7 pm> than other tellurite glasses, with approximately the same figures of merit. We
obtained an 80% diffraction loss and a 65% diffraction efficiency at a 633 nm wavelength, using
an 80 MHz 15 X 1 mm piezoelectric transducer with 1.5 W of rf power. This glass is able to
convert to fiber form, thus allowing efficient in-fiber acousto-optic modulators to be built.
1. INTRODUCTION
Materials particularly suitable for the construction of
acousto-optic (AO) devices exhibit high acousto-optic in-
teraction efficiency, low acoustic loss, high damage thresh-
old, small temperature effects, and good optical quality.’ In
addition, they should be readily available and easy to fab-
ricate, lap, and polish. Many crystalline materials exhibit
large acousto-optic figures of merit; however, the difficulty
in preparing monocrystals in large sizes and quantities lim-
its their suitability for A0 devices. Glasses, on the other
hand, are isotropic by nature and easy to fabricate in large
quantities with good optical quality, and it is usually pos-
sible to modify their transparency region without a signif-
icant sacrifice to their acousto-optic properties. Interest in
tellurite glasses as acousto-optic materials originated from
the fact that TeO,, the mother material of these glasses,
exhibits relatively high acousto-optic figures of merit’ and
is transparent in the range of 0.33-5.0 pm. A number of
tellurite glasses have been investigated in the past,213 how-
ever, their practical transparency region was reduced to
0.45-2.7 ,um. Here, we report on acousto-optic modulators
made of a new glass fabricated at Southampton University.
The glass composition is 3TeO,-2ZnC1, and it exhibits
transparency in the 0.42-5.7 ,um range (Fig. l), a range
wider than other reported2P3 A0 tellurite glasses. Our in-
terest in this glass system is mainly because we were able to
draw4 an optical fiber from it that can transmit in the
infrared range. Therefore, this glass will have a potential
use for building in-fiber acousto-optic modulators. Such
modulators have the advantage of low insertion loss, high
damage threshold, and negligible back reflection. They
have applications, for example, in fiber sensors, and in Q
switching and mode-locking of fiber lasers.
II. EXPERIMENT
The chlorotellurite (CLT) glass composition we used
is 3Te02-2ZnC1, and was fabricated in our laboratories.4
The Te02-ZnCl, glass with a TeO,/ZnCl, ratio of 1.5 was
prepared from dry mixtures of tellurium dioxide (reagent
grade, Aldrich chemicals, UK) and zinc chloride (analar
grade, Aldrich). Quantities of the TeO, and ZnCl, pow-
ders were mixed in platinum or alumina crucibles with a
capacity of about 100 ml. Owing to the hygroscopic nature
of ZnCl,, it was weighed first and then covered with the
TeO, to prevent as much water absorption as possible. The
mixture was then mixed well by mounting on a rotating
lathe for about 30 min. The mix was subsequently trans-
ferred to an electrically heated furnace preheated to 500 “C
for 2 h to further dry the mix. The furnace temperature
was then raised to 800 “C for glass melting and covered
with a lid to reduce the vaporization. The melt ‘was equil-
ibrated for approximately 20 min after melting, with fre-
quent shaking to achieve a bubble-free state. The melt was
then cast into a mild-steel mold with internal dimensions
approximately 50 mm long X 30 mm wide X 10 mm
thick. Immediately after solidification, the mold was re-
moved and the glass sample was put into a muffle furnace
at 260 “C and annealed for 1 h, then allowed to cool over-
night.
The sample used to build the acousto-optic modulator
(AOM) was polished face down to a A/4 flatness with
dimensions 16 X 10X 6 mm. The fabrication of an AOM
from this glass was performed by Gooch & Housego-UK,
Ltd. The piezoelectric transducer is a 42-,um-thick
LiNbO,, Y cut at 36” to;provide a compressional wave at
80 MHz. A lOOO-A-thick gold layer deposited on one of
the polished faces of the glass with the dimensions 6x 15
mm acts as the bottom electrode. On the opposite face a
metallic absorber is attached in order to ensure a traveling
wave operation. The transducer is bonded directly to the
gold by vacuum indium-cold-welding. A matching circuit
was connected to the transducer to minimize the reflected
rf power at resonance. The lateral dimensions of the trans-
ducer are L= 15 mm long and H= 1 mm wide. To char-
acterize the device, light from a HeNe laser at /2=633 nm
was launched into a single mode fiber and at the distal end
was recollimated using a quarter pitch GRIN lens. The
output from a single mode fiber was used for two reasons.
First, to obtain spatial filtering, which is found to improve
the diffraction efficiency. Second, to achieve a collimated
Gaussian laser beam with a small waist to demonstrate fast
switching under pulsed operation. The collimated beam
had a waist radius of tib= 145 ,um, measured by using the
knife edge technique.
J. Appl. Phys. 75 (l), 1 January 1994 0021-8979/94/75(1)/51 g/3/$8.00 0 1994 American Institute of Physics 519
Downloaded 18 Oct 2002 to 199.197.130.1. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/japo/japcr.jsp
O” 60
5
f 40
E 20
0 1 2 3 4 5 6 7 8
Wavelength (pm)
FIG. 1. Transmittance spectrum
of the
chlorotellurite
glass using a
2-mm-thick
sample. Fresnel
reflections were not ruled out
in
this spec-
trum.
III. RESULTS AND DISCUSSION
Figure 2 shows the variation of the diffraction loss and
the diffraction efficiency with rf power under cw operation
using a collimated HeNe laser beam at 633 nm with a beam
radius of 145 ,um. The diffraction loss is defined as the ratio
between the light intensity in the zeroth order beam when
the rf signal is ON to that when it is OFF. The diffraction
efficiency is the ratio between the light intensity diffracted
into the first order beam to the transmitted intensity when
the rf signal is OFF. The ditl’raction loss will, in general, be
higher than the diffraction efficiency because, apart from
the limiting case of operation in the Bragg regime, part of
the incident light is coupled into orders higher than the
first. In order to examine the speed of the device, the rf
signal was modulated to provide 500 ,us square pulses at a
0.6 kHz repetition rate and average r-f power of 0.45 W.
The acoustic velocity was estimated from a measurement
80
T
zi-
0.0 0.4 0.8 1.2 1.6
CW RF Power (W)
FIG. 2.
The
vartition
of the
measurd dif&&tion
loss and efficiency with FIG. 4. Variation of the measured deflection angle with the rf frequency
the rf
power.
The solid line is the fitted diffraction efficiency using Eq. under cw aperation using 0.45 W rf power. The solid curve is calculated
ill.
using the value of the acoustic velocity determined from Fig. 3.
6
8
t 2
2 4 6 8 10
Beam Position from the Transducer (mm)
FIG. 3. The delay time between the applied rf pulse and the optical
response as a function of the optical beam position from the transducer.
The rf pulse width is 500 ps, the repetition rate is 0.6 kHz, and the
average rf power is 0.45 W.
of the delay time between the rf pulse and the optical re-
sponse, as a function of the beam position from the trans-
ducer plane (Fig. 3 ) and found to be u= 3 108 f 93 m/s.
The value of v obtained by this technique is very accurate
(uncertainty is 3%) because the position of the optical
beam and the delay time are relative quantities measured
with respect to fixed points in space and time domains,
respectively. Under these conditions we measured a rise
time of 62 ns. The optical rise time 7, is given by Maydan’s’
formulae: TV,- 1.3wg’v, where mb= 145 pm is the beam ra-
dius. Using u=3108 m/s yields 7,=60 ns, in agreement
with the measured value. Figure 4 shows the variation of
the deflection angle 6 with the acoustic frequency V, exhib-
iting the linear behavior S=~Y/V. The straight line is cal-
30
q
measured
z
m
; 2O-
5,
P,
2
a
s 10 _
‘G
8
E
8
01 . , , , . ‘ . , . , ,
20 40 60 80 100 120 140
CW RF Frequency (MHz)
520 J. Appl. Phys., Vol. 75, No. 1, 1 January 1994 Abdulhalim et al.
Downloaded 18 Oct 2002 to 199.197.130.1. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/japo/japcr.jsp
TABLE 1. Physical parameters and estimated acousto-optic parameters
of CLT glass.
Optic31 transmission region [pm) 0.42-5.7
Refrxtive index (63.3 am) 2.0
Density (gem -‘) 4.4
Glass transition temperature I’C) 230
Acoustic attenuation at 80 hlHz (dB cm”“‘) < 1.0
Longitudinal acoustic velocity (m s- ‘) 3108
Young’s modulus (X10’” N m-‘) 4.25
Photoelastic constant P 0.09
MI I:*: IO-’ s rnz kg-’ ) 0.75
M2 ( >:. 1 or J 5 gs kg “.’ 1 ) 3.9
M, ( xlo-“l 2 m k&-g) 0.24
culated using the value of u, estimated from Fig. 3, and
exhibiting good agreement with the measured values of S
within the experimental error.
To determine the acousto-optic figures of merit, the
experimental results of the diffraction efficiency in the first
order Bragg diffraction (Fig. 2) were fitted to Gordon’s’T5
formulae:
y&l”[.4(P,)““] . (1)
Here A =n(PL1%$z/21$>0.‘/;1 with Iz being the optical
wavelength in vacuum, /3 is a factor that relates the acous-
tic power P, transferred to the A0 medium to the rf elec-
trical power Pti transferred to the transducer from the sig-
nal generator fi== PolPrf, and M,= (&‘/pv”) is the
second A0 figure of merit, where
p
is the photoelastic
constant and p is the density. The fitted curve in Fig. 2 was
obtained with A=0.761 w- which yields
&%fz-3.13 X 10-l’ ?/kg. The factor p was estimated to be
p~O.8 which then yields .&fz=3.9x lo-l5 ?/kg. From
this value of the acoustic velocity, the refractive index, and
the density of the CLT glass we estimated the effective
photoelastic constant and the other figures of merit
M, =n7p2/c)ll, and &fs = n7p2/pv’. The figures iw, , M,, and
&fj represent. scanning capability, diffraction efficiency,
and number of resolvable spots, respectively.’ To distin-
guish between the Bragg and Raman-Nath regimes, the
parameter Q=2rrL/nA” is usually introduced, where A is
the acoustic wavelength. The limit Q>l places the device
in the Bragg regime. In our case A=38.85 p.rn at 80 MHz,
which yields Q- 19.76. We observed a diffraction loss that
is substantially higher than the diffraction efficiency (into
the first order) (see Fig. 2). Operation not completely in
the Bragg regime will produce this effect, lowering the
maximum attainable value of the diffraction efficiency into
first order to less than 100% In our case, we believe that
residual inhomogeneities in the glass samples used are re-
sponsible for both a contribution to this effect. and a reduc-
tion in the measured figures of merit. Therefore the value
of &fz estimated by fitting the resu1t.s to Eq. ( 1) is the
minimum value. A more accurate estimate of the figures of
merit can be performed using the technique of Dixon and
Cohen.6
Table I summarizes the relevant parameters of the
CLT glass as an rZ0 material. The values of MI) &, and
Mj indicate that this glass is a highly promising A0 ma-
terial for modulators and deflectors which together with
the high acoustic velocity allows for fast switching and
scanning. It should be noted here that at high frequencies
the diffraction efficiency into the first order Bragg diffrac-
tion should improve. This is because the acoustic velocity
is high and therefore the acoustic wavelength A is larger
for the same frequency, so that if the transducer length is
insufficient for low frequencies (tens of MHz), additional
diffraction peaks can appear due to Raman-Nath type dif-
fraction. It should also be noted here that tellurite glasses””
exhibit low acoustic loss because of their short relaxation
time :P of the lattice phonons. The ultrasonic loss at fre
quencres up to few hundreds of MHz is proportional to rP
and increases quadratically with the frequency. It can be
described according to the Woodruff-Ehrenreich
equation?
a=4n-%Ty%%/3pu3. (2)
Here c is the specific heat, y is the mean Gruneisen param-
eter, and T is the temperature. We have not measured the
acoustic loss directly, but we did not observe any signifi-
cant decrease of the diffraction efficiency as the optical
beam position from the transducer was changed over a
distance of z 1 cm. This indicates that the CLT glass ex-
hibits low ultrasonic absorption at 80 MHz. This is not
surprising, however, since the glass network-forming oxide
TeOz, which exhibits low ultrasonic absorption ((r: < 1 dB/
cm> at 80 MHz, constitutes the major part of the compo-
sition. Low acoustic absorption means that higher fre-
quency operation of AOMs made of CLT glass is possible.
This increases the deflection angle, the bandwidth, the re-
solving power, and the diffraction efficiency.
In conclusion, we have found that the chlorotellurite
glass based on TeO,-ZnCl, systems can be used as an
acousto-optic material. The properties of this glass, com-
bined with its transparency in both the visible and the
midmfrared, nontoxicity, ease of fabrication, optical qual-
ity, and isotropic nature, make it attractive for acousto-
optic device applications.
ACKNOWLEDGMENTS
This work was supported by a U.K. government
(DTI) LINK project in collaboration with Gooch & Hou-
sego, Ltd* The Optoelectronics Research Centre is a U.K.
government, SERC sponsored, interdisciplinary research
centre. We would like to thank D. Moreau, G. Jones for
useful interactions and discussions.
’ L. N. Magdich and V. Ya. Molchanov,
Acousto-upfic Deices ard Their
.4pplcarions (Gordon & Breach, London, 1989).
“T. S. Izumiatani,
Opticai
Glass (American Institute of Physics, New
York, 19X6), pp~ 117-179.
-‘T. Yano, A. Fukumoto, and A. Watanabe, J. Appl. Phys. 42, 3674
(1971).
“J. Wang, Novel Multicomponent Glasses and Fibres for Fibre-optic
Device and Systems, Ph.D. thesis, University of Southampton, March
1993.
5 D, Maydan. IEEE J. Quark Electron. QE-6, 223 ( 1970).
“R. W. Dixon and M. G. Cohen, Appl. Phys. Lett. 8, 205 (1966).
‘T. 0, Woodruff and H. Ehrenreich, Phys. Rev. 123, 1553 (1961).
J. Appl. Phys., Vol. 75, No. 1, 1 January 1994 Abdulhalim et
al.
521
Downloaded 18 Oct 2002 to 199.197.130.1. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/japo/japcr.jsp
