772
J. Mater. Sci. Technol., Vol.25 No.6, 2009
CsCl Effected Ultrafast Third-order Optical Nonlinearities
of GeS2 -Sb2 S3 Chalcogenide Glasses
Hua Zhang1)† , Qiuhua Nie1) , Shixun Dai1) , Xiang Shen1) , Xunsi Wang2) and Xianghua Zhang1,2)
1) Faculty of Information Science and Engineering, The State Key Laboratory Base of Novel Functional Materials and
Preparation Science, Ningbo University, Ningbo 315211, China
2) Université de Rennes I, Rennes 35042, France
[Manuscript received July 17, 2008, in revised form October 14, 2008]
A series of alkali halide doped chalcohalide glasses (100−x)(0.9GeS2 -0.1Sb2 S3 )-xCsCl (x=5, 10, 15 and 20
mole fraction) were prepared. The absorption spectra and Raman scatting spectra of these glasses were
measured. The optical band gaps Eopt were obtained from ultraviolet absorption edges. Z-scan technique
was utilized to investigate the third-order nonlinear optical properties of GeS2 -Sb2 S3 -CsCl glasses. The value
of Eopt increases and the third-order optical nonlinearity decreases with increasing CsCl content. Decreasing
lone-pair electron and broadening the band-gap will provide less transition paths for nonlinear process, which
play a key role in ultrafast third-order nonlinear optical responses of these chalcohalide glasses.
KEY WORDS: Optical band gap; Microstructural units; Optical nonlinearity;
Chalcohalide glasses
1. Introduction
2. Experimental
Recently, chalcogenide glasses have attracted the
attention of many investigators due to their ability to
be used in infrared optics, photonic devices, reversible
optical recording and memory switching[1–3] . Chalcogenide glasses present exceptional infrared transparency, large nonlinear refractive index and low
phonon energies[4] . It has been reported that heavymetal ions such as Pb, Sb and Bi play an important
role in enhancing the refractive indices of glasses[5,6] .
Unfortunately, these glasses have low transparency in
the visual-IR (infrared) region. Therefore, the third
nonlinear property of glass samples with the lower absorption in visual-IR range can be easily investigated
by using Z-scan technique pumped around 800 nm.
It is suggested from Ge-Ga-Se system that adding alkali halide contents will improve the transmission of
the glasses[7] for the halide, situated closer to sulfur
non-bridging ions at the end of the glass network has
terminated effect and the large alkali cation plays the
role of modifiers.
Ternary chalcogenide Ge-Sb-S films have been intensively studied[8–10] . In order to improve transmission, cesium chlorine was added in glasses system.
Additionally, cesium chlorine can be a role of glassceramics nucleation in GeS2 -Sb2 S3 system[11–13] , prepared for making difference between chalcohalide
glasses and chalcogenide ceramics of third-order optical nonlinearity.
This paper presents the results of a systematic
study of the structure modification of chalcogenide
glasses in the Ge-Sb-S system with the progressive
increasing cesium chlorine. A correlation has been
established between the glass structure and the optical properties of the glasses. Raman scatting has
been used to examine the glass structure modification. Z-scan technique has also been used to study
the nonlinear optical properties of these new glasses.
Glassy samples of the (100−x)(0.9GeS2 0.1Sb2 S3 )-xCsCl (x=5, 10, 15 and 20 mole fraction) pseudo-ternary system were prepared by wellestablished melt-quenching technique. Elemental raw
materials of Ge, Sb, S and compound cesium chlorine
were carefully weighed to ±1 mg and transferred into
quartz ampoules within a N2 gas-filled glove box with
<1 ppm H2 O and O2 concentration. The quartz ampoules containing the raw materials were sealed under
vacuum with 10−3 Pa, which were then inserted into a
rocking furnace. Prior to sealing and melting, the ampoule and batch were pre-heated at 120◦ C for 2 h to
remove surface moisture from the quartz ampoule and
the batch raw materials. The quartz ampoules were
frequently agitated and maintained 12 h in the furnace in order to ensure homogeneity, then quenched
in water quickly to avoid crystallization. All samples
were annealed at 10◦ C below Tg (the glass transition
temperature) for 3 h to minimize inner tension induced by the quenching step and were slowly cooled
to room temperature. The sulfide samples were then
cut and optically polished to mirror smoothness with
a thickness of 1 mm for testing.
The absorption spectra of samples were recorded
in the range of 350–2700 nm using Perkin-Elmer
Lambda 950 UV-VIS-NIR spectrophotometer. Raman spectra of the samples were obtained using a Raman spectrometer (Advantage Nir) at the wavelength
of 785 nm (semiconductor laser) with power 70 mW.
The resolution of the Raman spectra was 1 cm−1 . Refractive indices were measured by SAIRON-SPA4000
Prism coupler. The ultrafast optical nonlinearity
of the glass samples were measured by Z-scan technique using a 76 MHz repetition rate mode-locked
Ti:sapphire laser (Coherent Mira 900-D) with 200 fs
pulse width. The schematic diagram of the experimental setup is depicted in Fig. 1. The laser radiation was split into two beams. One was detected by
detector 1 (D1) to monitor the fluctuation of the laser
energy, and the other was focused on the sample by a
† Corresponding author. Master; Tel.: +86 574 87600946; Fax:
+86 574 87600946; E-mail address: [email protected]
(H. Zhang).
J. Mater. Sci. Technol., Vol.25 No.6, 2009
773
Fig. 1 Z-scan experimental setup
Table 1 Glass composition, absorption edge, linear refractive index (n), the calculated nonlinear refractive index γ
(n2 ) and nonlinear absorption coefficient (β) of samples
Composition
Absorption edge
n
γ
β
n2
/mole fraction
/nm
/800 nm
/(10−14 cm2 /W)
/(cm/GW)
/(10−11 esu)
95(0.9GeS2 -0.1Sb2 S3 )-5CsCl
647
2.243
2.57
5.15
1.38
90(0.9GeS2 -0.1Sb2 S3 )-10CsCl
611
2.220
2.54
4.88
1.35
85(0.9GeS2 -0.1Sb2 S3 )-15CsCl
581
2.194
2.49
4.24
1.30
80(0.9GeS2 -0.1Sb2 S3 )-20CsCl
578
2.161
2.43
4.05
1.25
Note: n2 =(cn0 /40π)γ (10−14 cm2 /W) with c (m/s) the speed of light in vacuum and n0 the linear refraction index
1.0
Absorption / a.u.
0.8
0.6
0.4
0.2
500
600
700
800
900
1000
Wavelength / nm
Fig. 2 Absorption spectrum of 80(0.9GeS2 -0.1Sb2 S3 )20CsCl sample
lens with 150 cm focal length. As the sample moved
along a motor track near the focal point, the transmitted light changing with excitation intensity was
recorded by detector 2 (D2). D1 and D2 were connected to a dual-channel power meter, from which the
relative transmittance can be gained directly. Z-scan
of each sample was performed in the case of the openaperture (OA) and closed-aperture (CA). Measurements were performed at the wavelength of 800 nm.
All the measurements were taken at room temperature.
3. Results and Discussion
3.1 Optical absorption spectra
The optical linear absorption spectrum of
80(0.9GeS2 -0.1Sb2 S3 )-20CsCl glass in the visible and
near-IR region can be shown in Fig. 2. It indicates
that there is almost no absorption at 800 nm. The
absorption spectra of all samples are very similar to
each other except for minor shift of the absorption
edge with increasing CsCl content. As can be seen
from Table 1, the absorption-edges continuously shift
toward shorter wavelength with increasing amounts of
cesium chlorine, up to 20 mole fraction, which reaches
the limit of the glass formation. Correspondingly, the
color of the samples varies from red to orange following the increased content of cesium chlorine.
Corresponding to the transition between extended
in both valence and conduction bands, Taus and
Menth propose the following general relation to estimate optical band gap Eopt for the higher values of
the absorption coefficient[14,15] :
α(ω)~ω = B(~ω − Eopt )p
(1)
where B is a constant, p is an index that characterizes
the optical absorption process and is equal to 1/2, 2,
3/2 and 3 for direct allowed, indirect allowed, direct
forbidden and indirect forbidden transitions, respectively. For many glasses, the equation with p=1/2 and
p=2 is found to represent the experimental results.
The values of direct allowed optical band gap Edir−opt
and indirect allowed optical band gap Eindir−opt are
found by extrapolating the linear curve to zero absorption.
Plots of (α(ω)~ω)1/2 and (α(ω)~ω)2 as a function
of ~ω in eV unit of four glass samples can be shown in
Figs. 3 and 4. The values are found to be 1.927, 1.984,
2.106, 2.111 and 1.767, 1.868, 1.983, 2.011 which corresponds to cesium chlorine contents of 5%, 10%, 15%
and 20% (in mole fraction). It is shown in Fig. 5 that
introduction of cesium chlorine in 0.9GeS2 -0.1Sb2 S3
glasses increases the optical band gap Eopt . This can
be ascribed that Cl− ion has larger electronegation
774
J. Mater. Sci. Technol., Vol.25 No.6, 2009
4.0
2.15
2.15
E
dir-opt
h
(
1.95
1.90
1.90
1.85
1.85
E
Edir
1.80
2.0
1.5
2.00
1.95
E
2.5
2.05
2.00
/ eV
/ eV
3.0
2.10
2.05
indir-opt
3.5
/
/
2 )1 2
/ (cm
-
1
/
eV)1 2
indir-opt
2.10
1.80
-opt
1.75
1.6
5%
1.8
(
10%
h
15%
2 ) eV
/
2.0
20%
1.75
0
5
10
2.2
15
20
25
CsCl / mole fraction
/
Fig. 3 Plots of (αhω/2π)1/2 vs photon energy of
(100−x)(0.9GeS2 -0.1Sb2 S3 )-xCsCl glass samples
Fig. 5 Optical band gap energies (Edir−opt and
Eindir−opt ) as a function of CsCl content for
chalcohalide glasses
Intensity / a.u.
200
100
x=0.2
342
x=0.15
(
h
/
2 )2
/(cm
-
1
eV
)2
300
0
1.6
1.8
(
h
5% 10% 15%
2)
/
2.0
20%
x=0.1
x=0.05
2.2
/ eV
Fig. 4 Plots of (αhω/2π)2 vs photon energy of
(100−x)(0.9GeS2 -0.1Sb2 S3 )-xCsCl glass samples
than S2− , permitting to obtain of a better stabilization of the sulfur electron pairs. The incorporation of
Cl bonds to Ge or Sb in order to form non-bridging
atoms at the end of the glass network, while the large
Cs+ cation plays a role in modifiers, situated closer
to S− non-bridging ions. Therefore, the value of Eopt
increases and the visible absorption edge shifts to the
shorter wavelength with increasing CsCl contents.
3.2 Glass structure
In order to identify the covalent bonds inside glass
network and analyze the interrelationship between the
ultrafast optical nonlinearity and the optical band gap
through the introduction of CsCl in 0.9GeS2 -0.1Sb2 S3
sulfide glasses. Raman spectroscopy shows that for
all the samples, the local network is mainly in form of
[GeS4 ] tetrahedral unit with characteristic symmetrical stretching vibration band near 342 cm−1 and
[SbS3 ] pyramid unit plays miner role in the ration
between GeS2 and Sb2 S3 of 9 to 1[16] . The small
band around 230 cm−1 is due to vibration of ethanelike structure units [S3 Ge-GeS3 ], and the existence
of a Ge-Ge metal-metal band implies that the glassy
system becomes deficiency in sulfur according to the
chemistry prescription[16,17] . The shoulder at about
432 cm−1 originates from the vibration of two tetrahedral connected through bridging sulfur at the corner
or some multi-sulfur bonds[16,18] .
It can be seen from Fig. 6 that with increasing
CsCl contents into 0.9GeS2 -0.1Sb2 S3 network, the
432
270
230
200
250
300
350
400
450
Raman shift / cm
500
550
600
-1
Fig. 6 Raman spectra of (100−x)(0.9GeS2 -0.1Sb2 S3 )xCsCl glass samples
intensity at 270 cm−1 decreases obviously and shrinks
to nothing when x≥0.1, which is an only major change
in the detected Raman spectra. This is assigned to the
substituting of chlorine atoms to sulfur atoms in original [GeS4 ] tetrahedral that forms [S4−x GeClx ] (x=1,
2) tetrahedral and thus alleviates the S-deficiency of a
number of Ge-Ge bonds. In this process, because the
Cl− ion has only one coordinate bond, some bridging
sulfur chains have been broken and some non-bridging
sulfurs emerge in the glassy microstructure[18] . Cs+
ions can be dissolved into glass network as charge
compensators for non-bridging sulfurs.
3.3 Third-order optical nonlinearity
The typical Z-scan normalized transmittance
curves of the sample 80(0.9GeS2 -0.1Sb2 S3 )-20CsCl
can be seen from Fig. 7. Here the Z-scan measurement has been utilized at the power density
of I0 =2.55 GW/cm2 and the beam waist radius of
ω0 =32 µm. Each data point in Fig. 7 represents the
average of 20 shots. The solid liner is best-fit curves
obtained by the following equations for OA and CA,
respectively[19,20]
βI0 Leff
TOA (z) = 1 − √
2 2(1 + x2 )
(2)
4x(∆Φ0 )(1 − S)0.25
TCA (z) ∼
=1+
(1 + x2 )(9 + x2 )
(3)
775
J. Mater. Sci. Technol., Vol.25 No.6, 2009
1.2
Normalized transmittance
(a)
1.0
0.8
0.6
0.4
Experiment
Fitting
-15
-10
-5
0
5
Z / nm
10
15
pair from his outer orbit into the network, while Ge
atom has no lone pair. Therefore adding CsCl into
the network will decrease the density of the lone pair,
which contributes significantly to the polarizability of
the glass. The lone pair provides non-bonding levels
in the energy diagram; locating between the bonding and anti-bonding levels. Decreasing lone pair and
broaden the band-gap provide less transition paths for
nonlinear process. This mechanism plays a key role
in decreasing nonlinear polarizability. According to
the previous report[11] , excessive S–S covalent bond
can enhance the nonlinear response, so this is another
mechanism for increasing CsCl to decrease the nonlinear optical properties apart from the [Sb2 S3 ].
4. Conclusion
Normalized transmittance
(b)
1.5
1.0
0.5
Experiment
Fitting
0.0
-15
-10
-5
0
Z / mm
5
10
15
Fig. 7 (a) Open aperture Z-scan experiment data
and fitting curve of 80(0.9GeS2 -0.1Sb2 S3 )-20CsCl
(λ=800 nm, S=1), (b) the closed aperture Z-scan
experiment data and fitting curve of 80(0.9GeS2 0.1Sb2 S3 )-20CsCl (λ=800 nm, S=0.081)
where β is nonlinear absorption coefficient.
Leff is the sample effective length defined as
Leff =[1−exp(−αL)]/α (L:
the sample length).
x=z/z0 , where z is the distance and z0 is the diffraction length of the beam defined as z0 =πω02 /λ.
S is the aperture linear transmission. ∆ϕ is the
time-averaged peak-on axis phase change defined
as ∆ϕ=2πγI0 Leff /λ. α is the linear absorption
coefficient. When |∆ϕ0 | ≤ π, the ∆ϕ is approximately calculated by ∆Tp−v ≈0.406(1−S)0.25 ∆ϕ0 ,
∆Tv =βI0 Leff /2; ∆Tp−v is defined as the difference
between the normalized peak and valley transmittances in CA curve and ∆Tv is defined as valley
transmittances in OA curve. From the fitting cures,
we can get the ∆Tp−v and ∆Tv , then calculate the
nonlinear absorption coefficient β and the nonlinear
refraction index γ.
In order to determine the nonlinear refraction in
the presence of nonlinear absorption, the separation
and evaluation process have been performed: the
value of closed aperture (S<1) normalized Z-scan divide into one open aperture (S=1). According to
above calculation, the nonlinear refractive index γ
(n2 ) and nonlinear absorption coefficient β of glass
system can be shown in Table 1. It can be seen that
γ (n2 ) decreases with increasing CsCl contents, still
larger than heavy-metal oxide glasses[21,22] .
As we known, adding heavy metal content can increase nonlinear polarizability. Therefore in 0.9GeS2 0.1Sb2 S3 network, Sb atoms will improve the nonlinearity properties. Sb atom provides additional lone
The direct allowed optical band gaps Edir−opt and
indirect allowed optical band gaps Eindir−opt were obtained by extrapolating the linear portions to zero
absorption from plot of (α(ω)~ω)1/2 and (α(ω)~ω)2
as a function of ~ω in eV. The values of Eopt become
larger since the increasing CsCl content can increase
Cl− ions, which have larger electronegative than S2−
ions and can obtain the better stabilization of the sulfur electron pairs. Raman spectroscopy shows that
there were much more differences in glass structure
expect the intensity at 270 cm−1 . Decreasing lone
pair and broaden the band-gap provided less transition paths for nonlinear process, and played a key role
in ultrafast third-order nonlinear optical responses of
these chalcohalide glasses.
Acknowledgements
This work was supported by the National Basic Research Program of China (“973 Project”, (No.
2006CB708607), the Science and Technology Department
of Zhejiang Province, China (No. 2006C11127), the Science and Technology Foundation of Ningbo, China (No.
2006B100068) and the Natural Science Foundation of
Ningbo, China (Nos. 2006A610056, 2007A610004). This
work was also sponsored by K.C.Wong Magna Fund in
Ningbo University, China.
REFERENCES
[1] V.S. Vassilev and S.V. Boycheva: Talanta, 2005,
67(1), 20.
[2 ] C. Mihesan, S. Gurlui, M. Ziskind, B. Chazallon, G.
Martinelli, H. Zeghlache, M. Guignard, V. Nazabal,
F. Smektala and C. Focsa: Appl. Surf. Sci., 2005,
248(1-4), 224.
[3 ] A. Ganjoo, H. Jain, C. Yu, R. Song, J.V. Ryan, J. Irudayaraj, Y.J. Ding and C.G. Pantano: J. Non-Cryst.
Solids, 2006, 352(6-7), 584.
[4] H. Nasu, K. Kubodera, M. Kobayashi, M. Nakamura
and K. Kamiya: J. Am. Ceram. Soc., 1990, 73, 1974.
[5 ] N. Sugimoto, H. Kanbara, S. Fujiwara, K. Tanaka and
K. Hirao: Opt. Lett., 1996, 21(20), 1637.
[6] H.P. Li, B. Liu, C.H. Kam, Y.L. Lam, W.X. Que, L.M.
Gan, C.H. Chew and G.Q. Xu: Opt. Mater, 2000,
14(4), 321.
[7 ] L. Calvez, P. Lucas, M. Rozé, H.L. Ma, J. Lucas and
X.H. Zhang: Appl. Phys. A, 2007, 89(1), 183.
[8 ] K. Christova, A. Manov, V. Pamukchieva, A.G.
Fitzgerald and L. Jiang: J. Non-Cryst. Solids, 2003,
325(1-3), 142.
776
J. Mater. Sci. Technol., Vol.25 No.6, 2009
[9 ] EI-Sayed M. Farg: Opt. Laser Technol., 2006, 38, 14.
[10] E. Vateva and E. Savova: J. Non-Cryst. Solids, 1995,
192-193, 145.
[11] S.Z. Zhu, H.L. Ma, M. Matecki, X.H. Zhang, J.L.
Adam and J. Lucas: J. Non-Cryst. Solids, 2005,
351(40-42), 3309.
[12] X.H. Zhang, H.L. Ma and J. Lucas: J. Non-Cryst.
Solids, 2004, 337(2), 130.
[13] H. Ma, L. Calvez, B. Bureau, M.L. Floch and X.H.
Zhang: J. Phys. Chem. Solids, 2007, 68(5-6), 968.
[14] S.K.J. Al-Ani, C.A. Hogarth and R.A. EI-Malawany:
J. Mater. Sci., 1985, 20(2), 661.
[15] G. Saffarini, J.M. Saiter and H. Schmitt: Opt. Mater.,
2007, 29(9), 1143.
[16] I.P. Kotsalas, D. Papadimitriou, C. Raptis, M. Vlcek
and M. Frumar: J. Non-Cryst. Solids, 1998, 226(1-2),
85.
[17] S.S. Chu, F.M. Li, H.Z. Tao, H. Yang, S.F. Wang, C.G.
Lin, X.J. Zhao and Q.H. Gong: Opt. Mater., 2008.
[18] L. Petit, N. Carlie, F. Adamietz, M. Couzi, V. Rodriguez and K.C. Richardson: Mater. Chem. Phys.,
2006, 97(1), 64.
[19] M. Sheik-Bahae, A.A. Said, T.H. Wei, D.J. Hagan and
E.W. Van Stryland: IEEE J. Quant. Electron., 1990,
26(4), 760.
[20] T. Xia, D.J. Hagan, M. Sheik-Bahae and E.W. Van
Stryland: Opt. Lett., 1994, 19(5), 317.
[21] Y.F. Chen, Q.H. Nie, T.F. Xu, S.X. Dai, X.S. Wang
and X. Shen: J. Non-Cryst. Solids, 2008, 354(29),
3468.
[22] T. Hasegawa, T. Nagashima and N. Sugimoto: Opt.
Commun., 2005, 250(4-6), 411.