Characterization and properties of infrared NLO crystals: AGeX
3
(A ¼ Rb, Cs; X ¼ Cl, Br)
Zhi-Guang Lin
a,, Li-Chuan Tang
b, Chang-Pin Chou
a aDepartment of Mechanical Engineering, National Chiao Tung University, Hsinchu 305, Taiwan, ROC
bDepartment of Electrical Engineering, Chung-Cheng Institute of Technology, National Defense University, Taoyuan 353, Taiwan, ROC
a r t i c l e
i n f o
Available online 13 March 2008 PACS:
61.50.Nw 61.66.Fn Keywords:
A1. Optical characterization A1. Solid solutions A2. Growth from solutions B1. Halides
B2. Nonlinear optic materials
a b s t r a c t
Innovative infrared nonlinear optical crystals AGeX3were synthesized. Their powder X-ray diffraction patterns demonstrated that they had rhombohedral structures with R3m (no. 160) space group symmetry (except RbGeBr3). Their structural distortion increased with Br content, but decreased with Rb content. The Kurtz powder methods revealed that the nonlinear optical efficiency of CsGeBr3is 9.64 times that of rhombohedral CsGeCl3and 28.29 times that of KDP; most importantly, AGeX3is phase-matchable (except RbGeBr3).
&2008 Elsevier B.V. All rights reserved.
1. Introduction
Second-order nonlinear optical (NLO) materials played a key role in such optical fields as laser frequency conversion and optical parametric oscillation/amplification (OPO/OPA) [1,2]. For inorganic second-order NLO materials, several crystals used in ultraviolet (UV) and visible regions were proposed in the past two decades, such as KH2PO4(KDP), KTiOPO4(KTP), b-BaB2O4(BBO),
LiB3O5(LBO). But in the infrared (IR) region the current materials,
such as AgGaSe2, ZnGeP2, are not good enough for applications
mainly due to their low laser damage threshold, as their band gaps were smaller than 1.5 eV. Therefore, the search for new NLO crystals with excellent properties, especially a high damage threshold, has become a key area of research in NLO material science and laser technology[3].
Cesium trihalometalates CsMX3of group-IV elements (M ¼ Ge,
Sn, Pb; X ¼ Cl, Br, I ) have M2+(ns2) cations and crystallize in
perovskite variants [4–10]. Under ambient conditions CsMX3
adopts a rhombohedral structure, an attractive nonlinear crystal, which is applied in the infrared region. Since the optical damage threshold and the range of transparency of materials are related to the magnitude of the band gap, and the optical nonlinearity is inversely proportional to the cube of the band gap [11]. The linear and NLO properties of CsGe(BrxCl1x)3
and (RbyCs1y)GeBr3 can be adjusted by varying the alloy
composition to satisfy the demand for specific applications. This investigation presents a method for synthesizing crystals and measuring the optical properties for each composition. Nonlinear coefficients of CsGe(BrxCl1x)3, x ¼ 0;14;24;34;1 and y ¼ 0;14;24;34;1are
also considered to reveal the potential of these crystals in NLO applications.
2. Experimental details
2.1. Synthesis
The synthetic procedure was modified from that of Gu et al.
[12–14]. Chritensen and Rasmussen[15]and Tananaev et al.[16]
used different synthetic methods, but their methods seemed complex and offered poor productivity. In this work, H3PO2(50%)
was loaded with HBr (48%), HCl (37%) and GeO2 (99.999%) into
a 250 ml beaker, and then heated to 95 1C. The solution was vigorously mixed for 5 hours and then cooled to room temperature. After the precipitate was removed, CsBr (99.9%) and RbBr (99.8%) were added and the temperature was raised to the boiling point; the mixture was then naturally cooled to room temperature. A light yellow precipitate was formed. The reaction equations were as follows:
H3PO2þ6xHBr þ 6ð1 xÞHCl þ 2GeO2
¼H3PO4þ2HGeðBrxCl1xÞ3þ2H2O
Contents lists available atScienceDirect
journal homepage:www.elsevier.com/locate/jcrysgro
Journal of Crystal Growth
0022-0248/$ - see front matter & 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jcrysgro.2008.03.018
Corresponding author. Tel.: +886 3 5712121 55157; fax: +886 3 5720634. E-mail address:[email protected] (Z.-G. Lin).
then
HGeðBrxCl1xÞ3þCsBr ¼ CsGeðBrxCl1xÞ3k þ HBr
and
H3PO2þ6HBr þ 2GeO2¼H3PO4þ2HGeBr3þ2H2O
then
HGeBr3þyRbBr þ ð1 yÞCsBr
¼ ðRbyCs1yÞGeBr3k þ HBr
Recrystallization was performed by mixing the precipitate with 1:1 concentrated HX and alcohol solution to yield yellow crystals of CsGe(BrxCl1x)3 and (RbyCs1y)GeBr3. This approach was
performed 7 times to ensure no residual precursor remained. Then, the crystals were dried at 85 1C for 48 hours under a vacuum to prevent any effect of the deliquescence on them. The color of the precipitated product varied from yellow to white as soon as the substituted ratio, x, changed from 1 to 0.
2.2. Physical measurements
Rhombohedral crystals were synthesized and sieved into particles of different sizes to measure and analyze its structural and optical properties. The crystalline structures were observed using an X-ray diffractometer. The composition of all of the samples was measured by electron-probe X-ray microanalysis (EPMA). Raman spectra were obtained to determine the atomic vibration. The absorption edge was measured using a UV-visible spectrometer. Linear optical properties were measured using an ellipsometer. NLO properties were determined by making powder second-harmonic generation measurements.
3. Results and discussion
3.1. Composition and structural properties
Fig. 1 presents the composition of CsGe(BrxCl1x)3, (x ¼
0;1 4;
2 4;
3
4;1) and (RbyCs1y)GeBr3 (y ¼ 0;14; 2 4;
3
4;1) obtained from
EPMA measurements. These results indicate that those samples had a Cs to Ge ratio of almost 1:1 (for CGBC) and Ge to Br ratio of almost 1:3 (for RCGB). EPMA measurement qualitatively verified that chlorine atoms were successfully doped in the CsGeBr3 crystal and rubidium atoms were also
successfully doped in the CsGeBr3 crystal. Although it still
contained some impurities, they are all smaller than 1% (Omaxp0.52%, Pmaxp0.61%).
X-ray diffraction was employed to determine the structural parameters of all the crystals CsGe(BrxCl1x)3 and (RbyCs1y)
GeBr3. The synthesized crystals were crushed, ground and sieved.
X-ray diffractograms were obtained at room temperature using Cu-Ka radiation with Siemens D5000 equipment. FromFig. 2, the substitution-related diffraction peaks shifted gradually with substituted content for the anion substituted crystals. An extra CsBr crystal was used as an internal standard to determine the lattice parameters. The measured pattern was indexed and analyzed, as in the full-profile Rietveld refinement, using the nonprofit program PowderCell[17], which was developed by Kraus and Nolze. The structural parameters of CsGe(BrxCl1x)3 and
(RbyCs1y)GeBr3 were compared with those of CsGeCl3 (R3m,
no.160), CsGeBr3(R3m, no.160) and RbGeBr3(Pn21a, no.33), which
were presented in JCPDS. The peak-splitting caused by structural non-centrosymmetry, occurs mainly from 2y ¼ 151 to 351 (Fig. 2). The X-ray diffraction peaks shifted gradually as the substituted composition changed. Certain strong diffraction peaks were observed at 2y ¼ 31.761, 27.661, 26.861, 22.601, 22.101, 15.761 in CsGeBr3. These diffraction patterns were compared with those
with JCPDS and were indexed (2 0 0), (11¯1), (111), (11¯ 0), (11 0) and (1 0 0) planes, respectively. The XRD patterns also verify that CsGe(BrxCl1x)3 (x ¼ 0;14;24;34;1) and (RbyCs1y)GeBr3 (y ¼ 0;14;24;34)
crystallized in the non-centrosymmetric[18]rhombohedral space group R3m.
In an ideal perovskite structure, the cell parameters were a ¼ b ¼ c and a ¼ b ¼ g ¼ 901 with cubic space group Pm ¯3m (no. 221). Examples are the high-temperature phase of cubic CsGeCl3
and CsGeBr3 [18–22]: the cell parameters of cubic CsGeBr3 are
a ¼ b ¼ c ¼ 5.362 A˚ and a ¼ b ¼ g ¼ 901. The cell edges of rhombo-hedral (room temperature phase) CsGeBr3were longer than those of
cubic (high-temperature) phase, and the cell angles of rhombohe-dral (room temperature phase) CsGeBr3were slightly smaller than
901. The structural distortion contributes to the optical nonlinearity of CsGeBr3. Structural parameters from Fig. 3 indicated that the
lattice constant increased with Br content and the cell angle became smaller. Therefore, the structural distortion of CsGe(BrxCl1x)3
(R3m) increases with Br content. In contrast, the lattice constant became smaller as the Rb content increased while the cell angle became larger. Hence, the structure of (RbyCs1y)GeBr3 gradually
becomes centrosymmetric as Rb content increases.
3.2. Linear optical properties
Fig. 4shows the absorption spectra obtained at room tempera-ture using CsGe(BrxCl1x)3 (x ¼ 0;14;42;34;1) and (RbyCs1y)GeBr3
(y ¼ 0;1 4;
2 4;
3
4;1) crystals in the UV-visible light range. Thin plates
(E500 mm) of CsGe(BrxCl1x)3and (RbyCs1y)GeBr3were used to
make the band gap measurements. The recorded curves can be approximated as straight lines for a2 against hn, where a is the
absorption coefficient and hn is the photon energy. The straight-line approximation is applied to the rapidly increasing portions of the curves in Fig. 4. Hence, the fundamental absorption edge is described by a ¼ A(hn– Eg)1/2, where A is a constant and the band
gap Egcan be determined the points of intersection of the straight
lines with the abscissa. This dependence corresponds to directly
60 40 20 0 atom (%) CGC Br14 Br24 Br34 CGB Rb14 Rb24 Rb34 RGB Specimens 0 20 40 60 Cl Rb Br Ge P Cs O
Fig. 1. EPMA measurements, CsGe(BrxCl1x)3: x ¼ 0;14;42;34;1 and (RbyCs1y)GeBr3:y ¼
0;1 4;24;34;1.
allowed electronic transitions[23]. In the inset inFig. 4, the band gap values are plotted versus substituted content. Although the absorption edge declines from 3.43 to 2.38 eV as the bromine content (x ¼ 0–1), the absorption edge remains fixed for y ¼ 0–3
4.
In the Raman scattering measurement, the CsGe(BrxCl1x)3and
(RbyCs1y)GeBr3 samples were illuminated at room temperature
using an argon ion laser at 488 nm at an average power of 30 mW.
Fig. 5plots the results. The Raman peaks shift as the Br content, which fact is consistent with the fact that the phonon frequency is inversely proportional to the square root of the mass of GeX3
(X ¼ Cl, Br). Our data agree closely with the Raman spectra of CsGeCl3and CsGeBr3proposed by Thiele et al.[21]. The strongest
Raman peaks are associated with the A1mode. Raman peaks at
264–292 cm–1 can be also attributed to the A
1 mode. Another
group of Raman peaks (from 159 to 164 cm–1) is associated with
the E mode. The other group of Raman peaks (from 207 to
237 cm–1) can be attributed to the A
1+E mode and the A1mode
for CsGe(BrxCl1x)3 and (RbyCs1y)GeBr3, respectively. Raman
peaks of CsGeBr3at 419.21 and 209.79 cm–1are associated with
the corresponding (50.5 cm–1) overtones. Raman peaks of CsGeCl 3
at 236.98 cm–1can be assigned to the corresponding (58.5 cm–1)
overtone. From Fig. 6, lines (2–5,7) are associated with the anion substitution. Line (5) is associated with bromine atoms while line (7) relates to only chlorine atoms. Lines (2–4) are related to Ge–X3 bonds. This result is consistent
with the effective-mass concept: the oscillation frequency is expected to increase as the Br content declines because the Br atom is heavier than Cl. Lines (1,6,8) are attributed to the oscillation between Cs+ and Ge(BrxCl1x)3–1 because they
are less influenced by the anion substitution. In line (6), Raman signals were changed from the anion vibration A1+E modes to the cation vibration A1modes, revealing that the
relative vibration between Cs+ and the anion cluster
Ge(BrxCl1x)3–1was transformed to the doubly degenerate
vibra-tion of Ge–Br bonds.
3.3. Nonlinear optical properties
Powder SHG measurements, which were made by Chen et al.
[24], were made herein using a modified Kurtz-NLO system[25]
with 1260 nm light. A mode-locked Cr4+: Forsterite femtosecond laser with a pulse duration of 50 fs, was used in all measure-ments. The Cr4+: Forsterite oscillator yields pulses with a typical
FWHM bandwidth of about 45 nm at a repetition rate of 76 MHz and an average power of 270 mW. Since the SHG efficiency of powders has been shown to depend strongly on particle size [25,26], crystals were ground and sieved (Newark Wire Cloth Company) into six particle-size ranges 19–37, 37–74, 74–105, 105–210, 210–420 and 420–840 mm. Crystalline KDP was also ground and sieved into the same particle-size ranges to support relevant comparison with known SHG materials. All of the powders were placed in separate capillary tubes. The
5.70
5.55
5.40
0 50 100 50 100
Lattice parameter (angstrom)
Br composition (mol%) Rb composition (mol%) 88 89 90 CsGe(BrxCl1-x)3 (RbyCs1-y)GeBr3
Cell Angle
(degree)
Fig. 3. Structural parameters of the NLO crystals CsGe(BrxCl1x)3 and
(RbyCs1y)GeBr3.
Dif
fraction intensity (arb. unit)
20 30 20 30
Diffration angle 2θ (deg)
CsGe(BrxCl1-x)3 (RbyCs1-y)GeBr3
x=1.0 x=0.78 x=0.52 x=0.27 x=0 y=1.0 y=0.77 y=0.52 y=0.26 y=0
SHG radiation (630 nm) was collected in transmission and detected using a photomultiplier tube (Oriel Instruments). The SHG signal was collected by a data-acquisition (DAQ) interface and was monitored using an analysis program on a personal computer.
Powder SHG measurements of sieved polycrystalline CsGe(BrxCl1x)3 and (RbyCs1y)GeBr3 (Fig. 7) demonstrated that
the SHG efficiencies of all the samples exceeded that of KDP. Moreover, all of them were phase-matchable, as was KDP, such
that as the particles became substantially larger than the coherence length of the crystal, and the collected SHG intensity no longer increased, saturating at a particular value. The saturated PSHG intensities were estimated from the transmission signals in particle of various sizes, indicating that the SHG responses were strengthened as the Br content increased, but decayed as the Rb content increased. The deff values were calculated
(by dKDP¼0.36 pm/V[10]) and are plotted inFig. 8. The effective
1.0
0.5 0 50 100
0.0
3 2 3
1 2
Photon energy (eV)
α
2 (10
10
m
-1)
Br composition (mol%) Rb composition (mol%) 2.5
3.0 3.5
Energy Bandgap (eV) Energy Bandgap (eV)
2.8 2.6 2.4 0 50 100 CsGeCl3 CsGe(Cl1/4Br3/4)3 CsGe(Cl2/4Br2/4)3 CsGe(Cl3/4Br1/4)3 Rb3/4Cs1/4GeBr3 Rb2/4Cs2/4GeBr3 Rb1/4Cs3/4GeBr3 CsGeBr3 CsGeBr3 RbGeBr3
Fig. 4. Absorption coefficient near the band edge of CsGe(BrxCl1x)3and (RbyCs1y)GeBr3plotted in coordinates a2and hn. The inset shows the substituted composition
dependence of Egobtained.
Raman intensity (arb.unit)
100 200 300 400 150 300 450
Raman shift (cm-1)
CsGe(BrxCl1-x)3 (RbyCs1-y)GeBr3
x=1.0 x=0.78 x=0.52 x=0.27 x=0 y=1.0 y=0.77 y=0.52 y=0.26 y=0
powder second-harmonic generation coefficients increased with Br content, but decreased as the Rb content increased. Some reasons exist for the significant SHG signals of rhombohedral CsGe(BrxCl1x)3and (RbyCs1y)GeBr3crystals. First, the structural
distortion and the off-center Ge ion in the unit cell contributed to the SHG responses. The XRD results indicated that the structural distortion increased with Br content but decreased with Rb content. And the cell angle became smaller as Br content increased but became larger as Rb content increased. The position of the B-site cation, Ge, moved closer to the cell corner as the Br increased. Then, the nonlinearity increased with Br content, but declined as Rb content increased. Second, the optical
nonlinearity is approximately inversely proportional to the cube of the band gap value[11]. Therefore, the band gap values fell and the NLO susceptibilities increased with the atomic weights of the halides.
4. Conclusions
The structural and optical properties of rhombohedral NLO crystals, CsGe(BrxCl1x)3 (x ¼ 0;14;24;34and 1) and (RbyCs1y)GeBr3
(x ¼ 0;1 4;
2 4and
3
4), have been investigated experimentally to reveal
the substitution effect. Based on the results, the linearly increasing x increases lattice constant and second-order NLO susceptibility; but reduces the cell angle and the band gap values. The linearly increasing y yields opposite results. Owing to the optical damage threshold and the range of transparency of materials are related to the magnitude of the band gap, and the optical nonlinearity is inversely proportional to the cube of the band gap[11], the properties of halides could be modulated by anion substitution, but not cation substitution.
500 overtone overtone overtone overtone A1 mode A1 mode A1 mode A1 mode A1+E mode E mode E mode 250 0 0 50 100 50 100 Raman shift (cm -1)
Br composition (mol%) Rb composition (mol%) (7) (6) (4) (3) (2) (1) (5) (8)
Fig. 6. Raman vibrational modes of CsGe(BrxCl1x)3(R3m) and (RbyCs1y)GeBr3
(R3m) crystals.
1.6x106
8.0x105
0.0
0.0 0.3 0.6
PSHG intensity (arb. unit)
CsGeBr3 Particle size (mm) CsGe(Br3/4Cl1/4)3 CsGe(Br2/4Cl2/4)3 CsGe(Br1/4Cl3/4)3 (Rb1/4Cs3/4)GeBr3 (Rb2/4Cs2/4)GeBr3 (Rb3/4Cs1/4)GeBr3 CsGeCl3 KDP
Fig. 7. The comparison of integrated powder second-harmonic generation intensity of nonlinear optical crystals KDP, CsGe(BrxCl1x)3and (RbyCs1y)GeBr3.
10 def f ( pm /v )
CsGe(BrxCl1-x)3 (RbyCs1-y)GeBr3
1
0 50 100 50 100
2 3 4
Br composition (mol%) Rb composition (mol%)
Energy Bandgap (eV)
Fig. 8. The effective powder second-harmonic generation coefficients of nonlinear optical crystals CsGe(BrxCl1x)3and (RbyCs1y)GeBr3and their energy band gaps.
Acknowledgments
The authors would like to thank the National Science Council of the Republic of China, Taiwan, for financially supporting this research under Contract no. NSC 95-2112-M-009-042.
References
[1] D.M. Burland, Chem. Rev. 94 (1994) 1.
[2] D.S. Chemla, J. Zyss, Nonlinear Optical Properties of Organic Molecules and Crystals, Academic Press, Orlando, 1987.
[3] V.G. Dmitriev, G.G. Gurzadyan, D.N. Nikogosyan, Handbook of Nonlinear Optical Crystals, third ed, Springer, Berlin, 1999.
[4] J. Zhang, PhD thesis, Department of Material Science, Wuhan University, 1995. [5] M.D. Ewbank, F. Cunningham, R. Borwick, M.J. Rosker, P. Gunter, CLEO’97: Summaries of Papers Presented at the Conference on Lasers and Electro-Optics, vol. 11, 1997, 462pp.
[6] J. Zhang, N. Su, C. Yang, J. Qin, N. Ye, B. Wu, C. Chen, Chem. Proc. SPIE 3556 (1998) 1.
[7] L.C. Tang, C.S. Chang, J.Y. Huang, J. Phys.: Condens. Matter 12 (2000) 9129. [8] P. Ren, J. Qin, C. Chen, Inorg. Chem. 42 (2003) 8.
[9] P. Ren, J. Qin, T. Liu, Y. Wu, C. Chen, Optic. Mater. 23 (2003) 331.
[10] L.C. Tang, J.Y. Huang, C.S. Chang, M.H. Lee, L.Q. Liu, J. Phys.: Condens. Matter 17 (2005) 7275.
[11] Y.R. Shen, The Principles of Nonlinear Optics, Wiley, New York, 2002. [12] Q. Gu, Q. Pan, W. Shi, X. Sun, C. Fang, Prog. Cryst. Growth Charact. Mater. 40
(2000) 89.
[13] Q. Gu, Q. Pan, X. Wu, W. Shi, C. Fang, J. Crystal Growth 212 (2000) 605. [14] Q. Gu, C. Fang, W. Shi, X. Wu, Q. Pan, J. Crystal Growth 225 (2001) 501. [15] A.N. Christensen, S.E. Rasmussen, Acta Chem. Scand. 19 (1965) 421. [16] I.V. Tananaev, D.F. Dzhurinskii, Y.N. Mikhailov, Zh. Neorgan. Khim. 9 (1964)
1570.
[17] W. Kraus, G. Nolze, J. Appl. Crystallogr. 29 (1996) 301.
[18] D.K. Seo, N. Gupta, M.H. Whangbo, H. Hillebrecht, G. Thiele, Inorg. Chem. 37 (1998) 407.
[19] U. Schwarz, H. Hillebrecht, M. Kaupp, K. Syassen, H.G. von Schnering, G. Thiele, J. Solid State Chem. 118 (1995) 20.
[20] U. Schwarz, F. Wagner, K. Syassen, H. Hillebrecht, Phys. Rev. B 53 (1996) 12545.
[21] G. Thiele, H.W. Rotter, K.D. Schmidt, Z. Anorg, Allg. Chem. 545 (1987) 148.
[22] G. Thiele, H.W. Rotter, K.D. Schmidt, Z. Anorg, Allg. Chem. 559 (1988) 7. [23] J.I. Pankove, Optical Processes in Semiconductors, Prentice-Hall, Englewood
Cliffs, NJ, 1971.
[24] W.K. Chen, C.M. Cheng, J.Y. Huang, W.F. Hsieh, T.Y. Tseng, J. Phys. Chem. Solids 61 (2000) 969.
[25] S.K. Kurtz, T.T. Perry, J. Appl. Phys. 39 (1968) 3798. [26] J.P. Dougherty, S.K. Kurtz, J. Appl. Crystallogr. 9 (1976) 145.
