Metal-cladding directly defined active integrated optical waveguide device based on erbium-containing polymer

Abstract

In this study, a metal-cladding directly defined active waveguide technique is proposed. Optical waveguide amplifiers and thermo-optic (TO) waveguide switches based on erbium-containing polymer are designed and fabricated using this technique. The Er organic complexes containing polymeric reactivity groups are synthesized and investigated by free radical copolymerization. Optical characteristics and thermal stabilities of the active polymer are analyzed. Relative optical gain of the amplifier at 1530 nm is obtained as 3.6 dB. The rise and fall response time of the TO switch applied by 300 Hz square-wave voltage is measured as 511.7 μs and 341.5 μs, respectively. The power-time product is 12.5 mW ms and the extinction ratio is about 20 dB. The technique is very suitable for realizing large-scale optoelectronic integrated circuits.

---

1. Introduction

Active integrated optical waveguide devices, which contain functions of optical amplifiers and thermo-optic switches, are key components for reconfigurable optical add-drop multiplexing (ROADM) modules in optical cross connect (OXC) networks, which have important applications in optical communication, phase array antennas, optical computers, and optical sensing systems.^1–5 Especially, active polymer waveguides could offer a unique highly integrated optical signal processing platform with benefits of low loss, compact, flexible design of optical properties, and cost-effective features for high-density integrated devices.^6–8 As a multi-functional material system, active polymers exhibit well-controlled refractive indices, highly flexible structures, and large thermo-optic (TO) and electro-optic (EO) coefficients,^9–14 which can be advantageous for reducing manufacturing costs and opening the possibility of monolithic integration with functional devices such as optical amplifiers and switches. Several techniques, such as reactive ion etching,¹⁵ light-induced writing,¹⁶ laser writing,¹⁷ poling-induced writing,¹⁸ strain-inducing¹⁹ and molding/embossing,²⁰ have been developed to define optical waveguides in active polymers. In this study, we demonstrate a novel waveguide defining method that uses a metal-cladding confinement structure to directly produce a multi-functional integrated photonic chip based on erbium-containing polymers. Compared with other optical waveguide fabricating methods,^15–20 the metal-cladding directly defined waveguide technique could provide a simple, rapid and controllable fabricating process for an active integrated photonic device with low-cost, good-performance and effective-yield. Furthermore, compared to other Er-doped polymers,²¹ as the active core layer, Er complexes are bonded into the copolymer, which is advantageous for the complexes to uniformly disperse into the crosslink polymer, avoiding phase separation. Optical absorption characteristics and thermal stability of the cross-linked polymer were obtained. Optical properties of the active waveguide device were analyzed, simulated, and measured. The technique is advantageous for realizing a flexible fabrication process for large-scale photonic multi-functional integrated circuits (PICs) and optimizing cross-connectors for OXC and OADM systems.

2. Experimental section

2.1. Waveguide material

2-Thenoyltrifluoroacetone (TTA), 1,10-phenanthroline monohydrate (Phen) and erbium oxide (Er₂O₃) were purchased from J&K, Da Mao Reagent (Tianjin, China), Jin Cheng Yuan Co., and Aldrich (USA), respectively. Methacrylic acid (HMA), methyl methacrylate (MMA), glycidyl methacrylate (GMA) and other reagents were purchased from SCR (Shanghai, China) and Aladdin. They were used without further purification. 2,2′-Azobisisobutyronitrile (AIBN), freshly recrystallized, was used as a free-radical initiator. Copolymers were obtained by free-radical polymerization of dilute monomer solutions. All of the solvents were used after purification according to conventional methods. A total of 0.01 mol (3.825 g) of Er₂O₃ was dissolved in 40 mL of hydrochloric acid (HCl, 35 wt%) under stirring at 100 °C. The product Er chloride hexahydrate (ErCl₃·6H₂O) was obtained after evaporating excess HCl, washing with deionized water, and drying under vacuum. Finally, the product was obtained as a pink crystal.

As shown in Fig. 1(a), TTA (0.889 g, 4 mmol) and Phen (0.396 g, 2 mmol) were dissolved in ethanol (30 mL, 95%) and then mixed with HMA liquid (0.207 g, 2.4 mmol) in a three necked flask. The obtained solution was stirred and purged under N₂ at 50 °C. The pH value was carefully adjusted to 7.0–7.5 by adding 1 M NaOH in ethanol solution. After 30 min, 30 mL of 66.67 mM ErCl₃/ethanol solution was added slowly. With the addition of ErCl₃, the precipitate increased. The pH value of the mixed solution should be maintained between 7.0 and 7.5. The obtained solution was purged under N₂ for 20 h. The resulting complex Er(TTA)₂(Phen)(MA) was separated by centrifugation and rinsed with ethanol several times. The product was dried under vacuum to preserve it.

Fig. 1 Synthetic routes to (a) complex ErTPM and (b) copolymer GETPM.

The molecular structure and synthesis process of the Er-containing copolymer are given in Fig. 1(b). Er(TTA)₂(Phen)(MA) (ErTPM) (0.9 g, 1 mmol), GMA (4.4 g, 31 mmol), MMA (3.1 g, 31 mmol) and AIBN (0.15 g, 0.91 mmol) were dissolved in N,N-dimethylformamide (DMF, 40 mL) with stirring under a nitrogen atmosphere. The resulting solution was then heated at 70 °C for 6 h. The abovementioned mixture was transferred under stirring to a 500 mL beaker containing 200 mL of methanol. The flocculent precipitate was separated by centrifugation. The resulting polymer was dissolved in tetrahydrofuran (THF) and purified by precipitation in methanol; this procedure was repeated three times. The product (Er³⁺, 2 wt%) was then dried under vacuum. The yield of the copolymer was more than 90%. The molecular weight (M_n) was 38 000 and polydispersity (PD) was 1.82.

The complex ErTPM is polymerized with MMA and GMA by free-radical polymerization of dilute monomer solutions. The copolymers poly(GMA-co-Er(TTA)₂(Phen)(MA)) (GETPM) could form a highly cross-linked epoxy matrix structure, which exhibits good chemical resistance and excellent processability. Specific synthesis method and properties of the copolymer are given as ref. 22. By introducing an organic ligand (methacrylic acid) with excellent polymerization activity, carboxyl groups can be provided as coordinating groups for Er³⁺. Methacrylic acid plays a dual role as an organic ligand and a compatibilizer and Er chelates can use olefinic double bonds to copolymerize with other monomers for effectively reducing phase separation between Er complexes and polymers. There is low absorption for the Er-containing crosslinking polymer at 1310 nm and 1550 nm wavelength bands based on the near-infrared absorption spectrums of the Er-containing polymer. T_g is the glass transition temperature of Er-containing polymer, measured as 142 °C by differential scanning calorimetry (DSC); T_d is the onset temperature for 5% weight loss of the polymer, measured as 303 °C by thermal gravity analysis (TGA). The morphology and surface uniformity of the cured Er-containing polymer are measured by atomic force microscopy (AFM) and the root-mean-square surface roughness is 0.295 nm. These characteristics of the polymer material are suitable to improve the performances of integrated optical waveguide devices, which are contributed for optical communications.

In addition to 2 wt% concentration of Er³⁺ in the cross-linkable copolymer, other samples with different mass fractions of 1 wt% and 2.5 wt% Er-containing polymers were prepared using a similar method and under the same experimental conditions. The typical emission spectra of 1 wt%, 2 wt% and 2.5 wt% concentration Er-containing polymers were investigated as a KBr pellet (1 wt%) at room temperature, and the characteristic emissions from the Er³⁺ excited state were observed, as shown in Fig. 2. The wavelength of excitation was 390 nm. The emission spectra exhibit characteristic emissions in the region of 430–600 nm after the transitions of 4F_5/2 → 4I_15/2, 4F_7/2 → 4I_15/2, and 4S_3/2 → 4I_15/2 from Er³⁺. It can be found that 2 wt% concentration Er-containing polymer emits stronger characteristic luminescence than 1 wt% and 2.5 wt% samples in solid states excited at 390 nm. The reason may be that at low concentrations, luminescence intensity increases with increasing Er³⁺ concentrations. Above the critical point, the luminescence intensity will decrease because of luminescence-quenching occurring within the aggregates of Er³⁺ in GETPMs. We chose GETPM with 2.0 wt% Er³⁺ as the target product with the optimum proportion because of its aggregation and strong luminescence intensity.

Fig. 2 Emission spectra for 1 wt%, 2 wt%, and 2.5 wt% Er-containing copolymers in solid states excited at 390 nm.

Fig. 3 shows the infrared emission spectrum of poly(GMA-co-Er(TTA)₂(Phen)(MA)) at room temperature with the excitation of a 480 nm laser diode. The 2 wt% concentration of Er³⁺ in the cross-linkable copolymers has a great effect on the emission peak at 1530 nm, which is assigned to the 4I_13/2 → 4I_15/2 transition of Er³⁺. It is conducive to realize a fixed infrared signal light power amplification with the role of pump light.

Fig. 3 NIR photoluminescence spectra for poly(GMA-co-Er(TTA)₂(Phen)(MA)) in solid state at a excitation wavelength of 480 nm (75 mW mm²).

2.2. Waveguide design and fabrication

Fig. 4(a) shows the cross-sectional structure of the metal-cladding defined waveguide. Thermally oxidized silicon wafers were used as substrates, with a 5 μm thick oxide layer acting as the lower cladding layer. The thickness of Er-containing polymer is 3 μm after spin-coating onto a silicon substrate and heating at 120 °C. Double parallel Au strips with 0.3 μm thickness and 10 μm width were formed as metal upper claddings by the deposition, photolithography and wet etching technique. The active optical waveguide structure is directly defined with an air upper cladding layer between the Au strips. The refractive indices of SiO₂, Er-containing polymer, Au and air are 1.450, 1.505, 0.559 + 11.5i and 1.00 at 1550 nm wavelength, respectively. The mode distribution calculated by software COMSOL multiphysics is shown as Fig. 4(b). The relative eigenvalue equations based on the effective index method²³ are defined as


where n₁ is the refractive index of the core ridge waveguide, n₂ is the refractive index of the bottom cladding, n₃ is the refractive index of the top cladding, a is the width of the ridge waveguide, h is the height of the ridge waveguide, b is the thickness of the ridge waveguide, and N₁ and N₂ are equivalent refractive indices of different regions, respectively.

Fig. 4 (a) Cross-sectional structure of the metal-cladding defined waveguide and (b) mode distribution calculated based on the effective index method.

The equivalent refractive index of both-side waveguides with metal upper cladding is 1.4791 − 1.91 × 10⁻⁵i, and that of the core waveguide with air upper cladding is 1.4823. The fundamental mode effective index n_eff is 1.4803 + 6.586729 × 10⁻⁶i. The imaginary part of the effective refractive index is so small that it is ignored in the design of waveguide devices.

The direct optical waveguide amplifier and 1 × 1 MMI TO waveguide switch are designed and fabricated by the metal-cladding directly defined active waveguide technique. As shown in Fig. 5, the purpose is to realize a flexible monolithic multi-functional waveguide chip integrated with amplifying and switching properties.

Fig. 5 Schematic configuration of a monolithic multi-functional waveguide chip integrated with amplifying and switching properties.

The detailed fabrication process of the metal-cladding directly defined active waveguide device is shown in Fig. 6. The thin film was then cured at 120 °C for 1 h to achieve an epoxy cross-linked network with 2-methylimidazole as the initiator and to remove any traces of solvent. The gold film was deposited onto the Er-containing polymer core layer by the vacuum evaporation technique. The deposition time was 1 min and the vacuum reached 1.3 × 10⁻³ Pa. Furthermore, the BP212 photoresist was spin-coated on the gold film and pre-baked at 85 °C for 20 min to remove any solvent. The gold claddings with self-electrode structure were directly patterned using BP212 photoresist by photolithography and development. The pattern exposure was performed at a 365 nm wavelength with a 350 mW Hg lamp powered by an ABM high resolution mask aligner and exposure system with output intensity of 20 mW cm⁻² and exposure time of 6 s. The BP212 photoresist on the exposure area is removed in 5‰ NaOH solution and the gold claddings with self-electrode structure were formed in I₂ (1 wt%) + KI (4 wt%) developer at room temperature. Finally, the BP212 photoresist on the gold cladding was also exposed and removed in 5‰ NaOH solution. The metal-cladding directly defined waveguide process for the active device based on the Er-copolymer film has the advantage of a small number of steps apt for realizing high reproducibility of the sample fabrication.

Fig. 6 Fabrication process for metal-cladding directly defined waveguide structure.

Fig. 7(a) displays the structural patterns of double gold claddings and waveguide region by a microscope (×1000). It shows that the parameters designed for the metal-cladding directly defined active waveguide can be well realized and the process enables precise control of the core size. Fig. 7(b) gives the surface morphology from the gold cladding measured by AFM. The thickness is about 60 nm and the surface roughness less than 1.5 nm.

Fig. 7 (a) Structural patterns of double gold claddings and waveguide region by a microscope (×1000) and (b) surface morphology from the gold cladding measured by AFM.

2.3. Device measurement and discussion

For measuring the optical gain of the polymer direct waveguide amplifier, a tunable laser source (Santec TSL-210) with wavelength ranges from 1510 nm to 1590 nm was used as the signal source and a 980 nm laser diode with a maximum output power of 400 mW was used as the pump source. The signal and the pump light were launched into the channel waveguides by a 980/1530 nm wavelength division multiplexing (WDM) coupler. The output light from the device was collected and coupled to an optical spectrum analyzer (OSA, ANDO AQ-6315A). The transmission loss of the waveguide with 3 μm width was measured as 2.5 dB cm⁻¹ by a cut-back method. Fig. 8 shows the relative gain with increasing pump power at 1530 nm signal wavelength and the near-field pattern of the device. For an input signal power of 1 mW, the relative gain of the waveguide amplifier increased with the increasing pump power. When the pump power was 100 mW, the maximum relative optical gain at 1530 nm was obtained as ∼3.6 dB in a 5 mm-long waveguide. The relative gain of the amplifier can realize loss compensation for integrated optical waveguide circuits.

Fig. 8 Relative gain with increasing pump power at 1530 nm signal wavelength and the near-field pattern of the device.

A comparison with key performance parameters of other reports^24–26 is shown in Table 1. It could be observed that the proposed polymer waveguide amplifier based on the poly(GMA-co-Er(TTA)₂(Phen)(MA)) waveguide material could achieve stable operation well with higher relative gain by the metal-cladding directly defined technique. Compared with directly doping Er organic complexes, the Er-containing polymer produced by free radical copolymerization can increase Er³⁺ concentration in the polymer matrix and prevents degradation of performances for the optical amplifier.

Table 1 Comparison with other published results for polymer waveguide amplifier^a

APWMs	WL (mm)	RG (dB)	Reference
a APWMs = active polymer waveguide materials; WL = waveguide length; RG = relative gain.
LaF3:Er,Yb–LaF3/SGHM	17	3.5	24
NaY78%F4:Yb20%Er2%/KMBR	16	7.7	25
β-NaLuF4:Yb^3+, Er^3+ NCs/SU-8	12	2.413	26
Poly(GMA-co-Er(TTA)2(Phen)(MA))	5	3.6	Present work

---

1 × 1 multimode interference (MMI) TO switches are also realized based on Er-containing copolymers by the metal-cladding defined technique. The structural diagram of a 1 × 1 MMI switch device is constructed, as shown in Fig. 9 (a). Based on the MMI self-imaging theory,²⁷ the length and width of the MMI waveguide region are designed as 1290 μm and 35 μm, respectively. A taper structure enables both ends of the MMI waveguide to be connected to the single-mode input and output waveguides so that the switch can transmit optical signals adiabatically without any optical loss. The taper waveguide was optimized and defined as a length of 150 μm and an end width of 10 μm. The width of the gold electrode heater is designed as 10 μm, the width of the input and output waveguide is designed as 5 μm and an actual core size of 3 × 5 μm² is obtained for the waveguide device. The TO coefficient of the Er-containing polymer was measured as −1.65 × 10⁻⁴ °C⁻¹. The thermal conductivity of the SiO₂ buffer layer and active polymer was 0.27 and 0.16 W m⁻¹ K⁻¹, respectively. Fig. 9(b) illustrates the optical field transmission RSoft simulation results of the switch characteristics before and after thermal modulation, respectively. A square-wave voltage was applied to the electrode heater of the actual MMI TO switch with two needle-like probes. The measured total resistance was 300 Ω. The output signal light from the switch was coupled into a photodiode detector and observed using an oscilloscope.

Fig. 9 (a) The structural of the 1 × 1 MMI TO switch device and (b) the optical field simulation of the switching characteristics before and after thermal modulation, respectively.

Fig. 10(a) shows the TO switch response observed by applying square-wave voltage at a frequency of 300 Hz. The rise and fall times were 511.7 and 341.5 μs, respectively, and the average switch on–off time was about 400 μs. Fig. 10(b) shows the channel output intensity versus power consumption of the optical switch at 1550 nm for E^x₀₀ mode. The extinction ratio of the TO switch was about 20 dB. The applied electric power as the switching power was 23.5 mW, and the corresponding temperature increase of the electrode was 15 K.

Fig. 10 (a) Switch responses and (b) actual channel output versus power consumption of the MMI TO switch.

Compared with the reported polymer 1 × 1 MMI TO waveguide device,^28–31 including our study, the extinction ratio was obtained mainly in the range of 20–30 dB. The other parameter values of the device length and the switching power-time product were put into perspective by comparison with the performance of the polymer MMI TO device published in the literature, given in Table 2. It can be observed that the proposed 1 × 1 MMI TO device could achieve stable operation well with a smaller size and a lower power-time product. The advantages of the performances from the overall device can be obviously noted. In addition, the metal-cladding directly defined active waveguide technique without an etched or diffused process is more convenient for realizing integrated photonic circuits.

Table 2 Comparison with other published results for polymer 1 × 1 TO MMI device

In × out	Length (mm)	Power-time product (mW ms)	Reference
1 × 1	10	203	28
1 × 1	6	31.2	29
1 × 1	15	550	30
1 × 1	0.65	202	31
1 × 1	3	12.5	Present work

---

For the stability of the active devices, there was no damage phenomenon found in the amplifier with a maximum 980 nm pump power of 400 mW. Two months later, the response characteristics of the MMI TO switch were re-tested. Under the same testing condition, there was no attenuation of amplitude or hysteresis of switch time. For the durability of the active devices, the steady operating states of the amplifier and MMI TO switch were maintained and continuously enhanced the working temperature in the range of 25–85 °C held for 20 hours. There was no significant effect on performances of the active devices during the various reliability tests. The reasons for these findings may be that the 3D network of Er crosslinking copolymer forms a protective barrier against thermal decomposition while ensuring good stability and durability.

3. Conclusions

In summary, we demonstrated that new types of optical waveguide amplifiers and MMI TO waveguide switches based on erbium-containing polymer are achieved using a novel metal-cladding directly defined technique. The technique could provide a simple, rapid and controllable fabricating process for multi-functional integrated photonic chips and could have potential for surface plasmon polariton (SPP) waveguide applications. The core copolymers GETPM could form a highly cross-linked epoxy matrix structure, which exhibits good chemical resistance and excellent processability. By introducing an organic ligand (methacrylic acid) with excellent polymerization activity, carboxyl groups can be provided as coordinating groups for Er³⁺. Methacrylic acid plays a dual role as an organic ligand and a compatibilizer, and Er chelates can use olefinic double bonds to copolymerize with other monomers for effectively reducing phase separation between Er complexes and polymers. The optical properties of the active polymer were analysed. T_g and T_d of the Er copolymer are measured as 142 °C and 303 °C, respectively. The GETPM with 2.0 wt% Er³⁺ is chosen as the target product with the optimum proportion because of its aggregation and strong luminescence intensity. Relative optical gain of the amplifier at 1530 nm was obtained as 3.6 dB in a 5 mm-long waveguide. The characteristic parameters of the MMI TO switch were carefully designed and simulated. The power-time product was 12.5 mW ms and the extinction ratio was about 20 dB. This technique is advantageous for improving cross-connectors for optoelectronic integrated circuits.

Acknowledgements

The authors gratefully acknowledge financial support from the National Natural Science Foundation of China (No. 61107019, 61475061, 61405070, 61177027, 61275033, 61205032, 61261130586), Ph.D. Programs Foundation of Ministry of Education of China (No. 20110061120054, 20130061120060), Science and Technology Development Plan of Jilin Province (No. 20130522151JH, 20140519006JH), The Fundamental Research Funds for the Central Universities (No. JCKY-QKJC08) and National Undergraduate Training Programs for Innovation and Entrepreneurship (No. 2015510624).

References

1. T. Segawa, S. Matsuo, T. Kakitsuka, Y. Shibata, T. Sato, Y. Kawaguchi, Y. Kondo and R. Takahashi, Opt. Express, 2010, 18(5), 4341–4345.
2. F. Bontempi, S. Pinna, N. Andriolli, A. Bogoni, X. J. M. Leijtens, J. Bolk and G. Contestabile, IEEE J. Quantum Electron., 2012, 48(11), 1453–1461.
3. W. Wang, H. Zhou, J. Yang, M. Wang and X. Jiang, Opt. Lett., 2012, 37(12), 2307–2309.
4. F. Bontempi, S. Faralli, N. Andriolli and G. Contestabile, IEEE Photonics Technol. Lett., 2013, 25(22), 2178–2181.
5. M. Khoder, G. Verschaffelt, R. Modeste Nguimdo, X. Leijtens, J. Bolk and J. Danckaert, Opt. Lett., 2013, 38(14), 2608–2610.
6. C. Chen, X. Niu, C. Han, Z. Shi, X. Wang, X. Sun, F. Wang, Z. Cui and D. Zhang, Opt. Express, 2014, 22(9), 10716–10727.
7. Y.-T. Han, J.-U. Shin, S.-H. Park, J.-K. Seo, H.-J. Lee, W.-Y. Hwang, H.-H. Park and Y. Baek, IEEE Photonics Technol. Lett., 2012, 24(19), 1757–1760.
8. C. Chen, F. Zhang, H. Wang, X. Sun, F. Wang, Z. Cui and D. Zhang, IEEE J. Quantum Electron., 2011, 47(7), 959–964.
9. N. Bamiedakis, J. Beals IV, R. V. Penty, I. H. White, J. V. DeGroot Jr and T. V. Clapp, IEEE J. Quantum Electron., 2009, 45(4), 415–424.
10. J. Liu, G. Xu, F. Liu, I. Kityk, X. Liua and Z. Zhen, RSC Adv., 2015, 5, 15784–15794.
11. W. Huang, L. Chen and L. Xuan, RSC Adv., 2014, 4, 38606–38613.
12. J. Luo, X.-H. Zhou and A. K. Y. Jen, J. Mater. Chem., 2009, 19, 7410–7424.
13. J. Wu, H. Xiao, L. Qiu, Z. Zhen, X. Liu and S. Bo, RSC Adv., 2014, 4, 49737–49744.
14. J. Wu, C. Peng, H. Xiao, S. Bo, L. Qiu, Z. Zhen and X. Liu, Dyes Pigm., 2014, 104, 15–23.
15. S.-K. Kim, H. Zhang, D. H. Chang, C. Zhang, C. Wang, W. H. Steier and H. R. Fetterman, IEEE Photonics Technol. Lett., 2003, 15(2), 218–220.
16. O. Sugihara, S. Yasuda, B. Cai, K. Komatsu and T. Kaino, Opt. Lett., 2008, 33(3), 294–296.
17. L. Eldada, C. Xu, K. M. T. Stengel, L. W. Shacklette and J. T. Yardley, J. Lightwave Technol., 1996, 14(7), 1704–1713.
18. S.-K. Kim, K. Geary, D. H. Chang, H. R. Fetterman, H. Zhang, C. Zhang, C. Wang and W. H. Steier, Electron. Lett., 2003, 39(9), 721–722.
19. S.-K. Kim, W. Yuan, K. Geary, Yu-C. Hung, H. R. Fetterman and D.-G. Lee, Appl. Phys. Lett., 2005, 87, 011107.
20. A. Rogner and H. Pannhoff, Optical Fiber Communication (OFC'94), Tech. Digs., San Jose, CA, USA, 1994, vol. 4, pp. 279–280.
21. Y. Zheng, Y. Zhou, G. Accorsi and N. Armaroli, J. Rare Earths, 2008, 26, 173–177.
22. D. Fan, X. Fei, J. Tian, L. Xu, X. Wang, S. Fan and Y. Wang, Polym. Chem., 2015, 6, 5430–5436.
23. C. Chen, X. Sun, F. Wang, F. Zhang, H. Wang, Z. Shi, Z. Cui and D. Zhang, IEEE J. Quantum Electron., 2012, 48(1), 61–66.
24. S. H. Bo, J. Hu, Z. Chen, Q. Wang, G. M. Xu, X. H. Liu and Z. Zhen, Appl. Phys. B, 2009, 97(3), 665–669.
25. K.-L. Lei, C.-F. Chow, K.-C. Tsang, E. N. Y. Lei, V. A. L. Roy, M. H. W. Lam, C. S. Lee, E. Y. B. Pun and J. Li, J. Mater. Chem., 2010, 20, 7526–7529.
26. Y. Wang, X. Guo, S. Liu, K. Zheng, G. Qin and W. Qin, J. Fluorine Chem., 2015, 175, 125–128.
27. J.-K. Hong and S.-S. Lee, IEEE Journal of Lightwave Technology, 2007, 25(5), 1264–1268.
28. Y.-O. Noh, C.-H. Lee, J.-M. Kim, W.-Y. Hwang, Y.-H. Won, H.-J. Lee, S.-G. Han and M.-C. Oh, Opt. Commun., 2004, 242, 533–540.
29. J. Xiaoqing, L. Xia, Z. Haifeng, Y. Jianyi, W. Minghua, W. Yuying and I. Shigeta, IEEE Photonics Technol. Lett., 2005, 17, 2361–2363.
30. T. K. Jin, C. Choon-Gi and S. Hee-Kyung, ETRI Journal, 2005, 27, 122–125.
31. A. Maese-Novo, Z. Zhang, G. Irmscher, A. Polatynski, T. Mueller, D. de Felipe, M. Kleinert, W. Brinker, C. Zawadzki and N. Keil, Appl. Opt., 2015, 54(3), 569–575.

---