In₂Te₃ thin films: a promising nonlinear optical material with tunable nonlinear absorption response

Abstract

A series of In₂Te₃ thin films with various thicknesses was prepared on fused quartz substrate using a radio-frequency magnetron sputtering method. After annealing, the surface morphology, structure and linear optical absorption of these films was investigated. By performing an open aperture Z-scan technique with femtosecond pulses under 800 nm and 400 nm wavelengths, both the sign and magnitude of nonlinear absorption coefficient can be deduced. The experimental results show that In₂Te₃ thin films exhibit a large and ultrafast third-order nonlinear optical absorption response. Interestingly enough, a turnover from saturation absorption to reverse saturation absorption is observed by adjusting the film's thickness. Herein, a three-level model is presented to determine the tuning behavior and the competition mechanism between ground state absorption and excited state absorption. Moreover, the ultrafast pump-probe experiment furnished evidence of a nonlinear absorption difference, which was mainly caused by the density of defect states.

---

Introduction

Since the discovery of the first laser in 1960, investigations of nonlinear optical absorption (NOA) materials have been developing rapidly due to their extensive applications in the field of civilian photoelectric devices as well as military defence systems. Over the past few decades a variety of applications based on NOA materials have been realised, such as multi-photon pumped lasing,¹ ultrashort pulse generation² and optical limiting. Some new emerging areas include super-resolution imaging,^3,4 passive optical diodes,⁵ and ultrafast optical communication.⁶ In the pursuit of new functional fields, therefore, finding and investigating advanced NOA materials is of great importance not only to understand the basic physical mechanism of nonlinear absorption but also to design versatile nonlinear optical devices.

The origins of nonlinear absorption vary widely; for example, NOA may be associated with two-photon absorption (TPA) and excited state absorption (ESA) in the femtosecond domain, and free carrier absorption in the picosecond domain. In general, two opposite types of NOA can be distinguished. One is where transmittance increases as input laser intensity increases; this is called saturation absorption (SA). The other is where transmittance diminishes as input laser intensity increases; this is called reverse saturation absorption (RSA).⁷ Under the current circumstances, most NOA devices are merely based on SA or RSA independently of each other because of the restriction of optical materials.⁸ In recent years, however, some studies have adopted a combination model where SA and RSA materials were applied simultaneously. Anand et al. reported a solid state all-optical diode with a thin SA layer (graphene) and a thin RSA layer (C₆₀) in tandem.⁵ Band et al. proposed an arrangement of a saturable absorber followed with a reverse saturable absorber to reduce the trailing edge of optical pulses.⁹ S. Roy and C. Yadav presented a detailed theoretical model of femtosecond all-optical parallel logic gates based on tunable saturable to reverse saturable absorption in graphene oxide thin films.¹⁰ In addition, the tuning from SA to RSA has been observed and investigated in CuPc-doped PMMA thin films and graphene oxide thin films,^11,12 which attribute the tuning to TPA-induced ESA. MoS₂ nanoflake array films and Au nanocubes are also found to be NOA-tunable because of the competition mechanism of ground state absorption and excited state absorption in these materials.^13,14 The combination of SA and RSA materials will undoubtedly open up new access to obtain distinctive nonlinear optical configurations.

As a member of the chalcogenide alloys, indium telluride (In₂Te₃) is a III–VI semiconductor with defect structure, in which the presence of defects can significantly enhance the diversity of NOA. The two phases existing in In₂Te₃ crystalline structure are the α-phase with anti-fluorite structure and β-phase with zinc blende structure. It has always been regarded as a type of ideal material for photodetectors,¹⁵ gas detectors,¹⁶ and thermoelectric power generators due to its outstanding optical and electrical properties,^17,18 showing versatility and tailoring characteristics. In spite of this, the overwhelming majority of reports on In₂Te₃ thin films are concentrate on linear optical and electrical applications.¹⁹ To our knowledge, there have been no descriptions in the literature of three-order nonlinearity to date. In this study, we investigate the NOA properties of In₂Te₃ thin films and an exhilarating reversal of NOA behavior found only by thickness control, which provides a simple design and cost-effective way to develop innovative NOA devices.

Experimental

Sample preparation

Four different thicknesses of In₂Te₃ thin films were deposited on fused quartz substrates by a radio-frequency (RF) magnetron sputtering method using a 99.99% In₂Te₃ target in a working pressure of 3 × 10⁻³ mbar. Substrates were maintained at room temperature during the sputtering process. The thickness of the films was controlled by varying deposition time, while the sputtering rate was kept constant. The deposition times were 200 s, 400 s, 600 s, and 800 s, with resulting thicknesses labelled as S1, S2, S3, S4, respectively. To obtain crystalline structures, all the films were annealed in flowing nitrogen ambient at 573 K for 30 minutes.

Characterization methods

A surface profiler (Veeco Dektak 150) was employed to determine the thickness of the annealed films. Atomic force microscopy (Shimadzu SPM-9500J3) was used to observe surface morphology. Absorption spectroscopic characterization was carried out using a double beam UV-VIS-NIR spectrophotometer (Shimadzu UV-3600), where a fused quartz substrate was set in a reference light path to offset the substrate's absorption. The films' structural analysis was carried out using an X-ray diffractometer (Bruker D8 ADVANCE), with a Cu-K_α (λ = 1.54056 Å) radiation source in a 2θ range of 10–70°.

Z-scan technique and its calibration

The single beam Z-scan technique was performed to detect the nonlinear optical absorption properties of In₂Te₃ thin films. Fig. 1 displays the 3D sketch map of our Z-scan system. The excitation pulses were generated by a Ti:sapphire regenerative amplifier system (Spectra Physics, Spitfire Ace) at 800 nm and 400 nm (frequency-double) wavelengths with 100 fs pulse duration and 1 kHz repetition rate. Its output power can be carefully controlled by two tunable attenuators. The transmitted beam from the beam splitter was directly recorded by a photodiode power detector, D1, (Newport 918D) to monitor pulse energy stability. The reflected beam from the beam splitter was focused by a convex lens (f = 300 mm) and then vertically irradiated onto the In₂Te₃ thin film samples, which were moved along the propagation direction using a stepping motor, whereas the nonlinear transmittance was recorded using a similar power detector, D2. The beam radius ω at the focal point was estimated to be about 32 μm by means of repeated knife-edge measurements. The Rayleigh length of Gaussian beam was calculated to be about 4 mm, much larger than the thickness of each sample and substrate, which was an essential requirement for validity of the Z-scan theory.²⁰ Our open-aperture (OA) mode Z-scan was achieved without the aperture in front of the detector D2. Carbon disulfide solution in a 1 mm thick quartz cuvette was adopted as the standard for Z-scan system calibration. Analysing the OA Z-scan data, the obtained value of nonlinear absorption coefficient of carbon disulfide solution is 6.2 × 10⁻¹¹ cm W⁻¹, which is very close to the result reported in the literature.²¹ This demonstrates validity of our Z-scan system.

Fig. 1 Schematic of Z-scan experimental setup, OA Z-scan can be achieved by removing the aperture in front of detector 2.

Results and discussion

Morphological and structural analysis

By scanning the height difference between the top film surface and the fused quartz substrate, the film thicknesses with different deposition times were measured to be 43 nm, 78 nm, 121 nm, and 165 nm, respectively, corresponding to S1, S2, S3 and S4. As shown in Fig. 2, surface morphology images denote that the crystallized In₂Te₃ films have a homogeneous surface. In thinner films, the surface is rather smooth and the nearly-rounded grains indicate that the films start crystallizing. The grain size is then increased and surface roughness increases quickly to about 1.68 nm and 1.96 nm for S3 and S4, respectively; the grains tend to agglomerate with rounded particles to form bigger grains, indicating better crystallization.

Fig. 2 (a–d) AFM images and thickness of In₂Te₃ thin films with sputtering times of 200 s, 400 s, 600 s and 800 s, respectively. The root mean square is used as the reference of surface roughness.

Presented in Fig. 3 are the XRD patterns of the deposited and annealed In₂Te₃ thin films. Lack of diffraction peaks in the as-deposited films reveals their amorphous nature. After the thermal treatment, the intensity of the diffraction peaks of annealed samples increases as the thickness increases, and also new diffraction peaks appear, implying the improvement of crystallinity for these films. The peak positions of XRD patterns show that all the annealed samples have zinc blende structures, corresponding to the β-In₂Te₃ phase. In addition, β-In₂Te₃ thin films exhibit preferred (111) and (220) orientations as previously reported.¹⁸

Fig. 3 XRD patterns of deposited In₂Te₃ thin films and annealed In₂Te₃ thin films with different thickness. The broader peaks from 10° to 30° come from the quartz substrate.

Linear optical properties and defects

To preliminarily understand the NOA mechanism, the linear optical absorption spectra of β-In₂Te₃ thin films were examined and are shown in Fig. 4; the films have some absorption at the excitation wavelength of 800 nm, corresponding to single photon absorption. Importantly, the films have more absorption at 400 nm indicating the probability of excited states absorption in two steps or genuine two-photon absorption.²² The absorption spectra of crystallized In₂Te₃ thin films have extended tails over a range from the visible to the near infrared, indicating the presence of different local levels in the forbidden gap. These local levels primarily originate from the amorphous state and crystal structure defects, such as vacancies, stacking faults and surface defects in In₂Te₃ nanostructures.^18,23 Particularly in S3 and S4, fluctuating trends can be observed in the visible and near-infrared region; clearly the concentration of these local defect states is very high and they can make absorption jump. It is usually believed that thermal treatment is a typical method to reinforce crystallization and release unsaturated defects of semiconductor thin films,²⁴ whereas in our experiments there still remain some defect states in the crystallized thin films due to the complexity of internal structural defects. Defects are usually considered to be unwanted in bulk materials since an imperfect structure may degrade device performance, but in the case of nanostructure materials the presence of defects can significantly enhance the diversity of performance and tailoring property. Concisely, after annealing, there still exist defect states in In₂Te₃ thin films, and the various optical absorption behaviors depend largely on the existence of these defect states.

Fig. 4 Absorption spectra of annealed In₂Te₃ thin films.

Moreover, the absorption spectra of samples are notably different and an influence of film thickness on the band gap can be expected. To determine the band gaps the Tauc plot method was applied. Plotting the dependence of (αhν)² vs. hν will give a straight line in the linear region and extrapolating the linear portion will give the value of optical band gaps for allowed direct transitions,²⁵ where α is the linear optical absorption coefficient and hν is the photon energy. As demonstrated in Fig. 5, it can be seen that band gaps decrease from 1.95 eV to 1.71 eV with increasing thickness. Despite being slightly larger, the band gaps obtained by Tauc plot method are comparable with the reported values since the presence of defects may somewhat affect the band gap structure.²⁶ Moreover, a similar band gap dependence on the thickness is also observed in other III–VI semiconductors, the band gap widening is attributed to the quantum-confinement effect.^27,28

Fig. 5 Extrapolation of Tauc plots of annealed In₂Te₃ thin films.

Nonlinear optical absorption

Nonlinear optical properties of the amorphous and crystallized In₂Te₃ thin films are investigated through the single beam OA mode Z-scan method with femtosecond laser pulses under the 800 nm wavelength. It should be noted that the fused quartz substrates were measured in advance and they had no evident nonlinear absorption response until the power density I₀ at the focal point reached about 3600 GW cm⁻². Therefore, the NOA response observed herein, is found to originate only from the In₂Te₃ thin films. During the measurements, low-intensity pulses were carefully chosen to avoid both high-order nonlinearity and nonlinear scattering. By replacing the carbon disulfide solution with In₂Te₃ thin films and other elements unchanged, as shown in Fig. 1, the OA Z-scan measurements on In₂Te₃ samples were performed at an incident power density of about 60 GW cm⁻². The results are depicted in Fig. 6(a) and (b), the dots represent recorded normalized data, whereas the curves are the best-fitting results. Curves comprise clear normalized transmittance peaks and valleys, indicating the presence of SA and RSA; NOA type evolves from SA into RSA as the film thickness increases. All samples still have a similar trend of NOA tuning characteristics after annealing, but with smaller modulation depths (peak heights or valley depths). All the Z-scan traces of In₂Te₃ thin films keep symmetry with respect to the focal point and repeatable nonlinear absorption reversal can be achieved, indicating that the resulting Z-scan curves originate from the intense light-induced nonlinear absorption rather than structure- or phase change-induced nonlinearity.²⁹

Fig. 6 OA Z-scan curves at I₀ = 60 GW cm⁻²: (a) as-deposited In₂Te₃ thin films and at 800 nm (b) annealed In₂Te₃ thin films at 800 nm. (c) Schematic of three-level model for 800 nm excitation wavelength, where σ_gr is the cross-section of ground states absorption and σ_ex the cross-section of excited states absorption. (d) As-deposited In₂Te₃ thin films and at 400 nm (e) annealed In₂Te₃ thin films at 400 nm and (f) schematic of three-level model for 400 nm excitation wavelength.

Generally, SA occurs only in semiconductors with band gaps smaller than the incident photon energy, resulting from the electrons' excitation from valence band to conduction band, and then Pauli-blocking induced bleaching effect.⁷ However, the laser pulses at 800 nm (at 1.56 eV i.e. smaller than the band gap) still induce SA for the thinner sample films, and RSA for the thicker ones. Comparing Fig. 6(a) and (b), it can be speculated that defect states provide a new channel for electrons to transfer from valance band to conduction band. After thermal treatment, part of the defect states removed by relaxation and the possibility of defect states-assisted transition declines, leading to smaller modulation depth. A three-level model is proposed to explain the NOA behaviors, as shown in Fig. 6(c); electrons transfer from valence band to defect band by absorbing excitation photons. These electrons were dynamically stranded in defect states, thus resulting in the SA response. Simultaneously, some of the electrons in the defect band may transfer to higher excited states located in the conduction band by absorbing spare photons i.e. excited state absorption (ESA).

The turnover from SA to RSA is usually considered as the result of competition between ground states absorption and excited states absorption. In thinner films (S1 and S2), the band gaps are larger and it is difficult for electrons to transfer from defect states to higher states in the conduction band; most electrons are stranded in the defect band, corresponding to SA response.³⁰ On the contrary, the narrower band gaps in thicker samples (S3 and S4) can easily trigger excited state absorption; SA behavior weakens and RSA behavior tends to gradually become dominant. During the whole process, reversal of NOA type can be realized through band gap tailoring. To corroborate the three-level model, we performed the 400 nm OA Z-scan experiment in the same input laser intensity. Unlike the 800 nm excited group, as shown in Fig. 6(d) and (e), the 400 nm excited group shows totally SA response for all the films. In Fig. 6(f), electrons can directly transfer from valence band to conduction band by absorption of an excitation photon (3.1 eV); the result is in line with our proposed mechanism.

Ultrafast optical pump-probe spectroscopy

For a better understanding of the abovementioned defect state levels and also their relaxation mechanism, a photon-induced transient absorption spectrum was measured through ultrafast pump-probe method, in which the In₂Te₃ thin films were excited at 800 nm wavelength using the Ti:sapphire laser and a relatively weaker probe pulse at 950 nm; a broad range of delay time is used to observe the relaxation dynamics of the excited electrons.

While exciting the samples at 800 nm, the electrons are pumped to the shallow defect levels below the band gaps (635–725 nm). Unlike the OA Z-scan experiment, the pump pulse intensity in the pump-probe test is much lower to avoid bleaching signal. The pump-probe result in Fig. 7 shows that there is a significant difference in electron relaxation behavior between the thinner samples (S1, S2) and the thicker ones (S3, S4). The bi-exponential function was chosen to fit the decay process given by the following equation:

Fig. 7 Decay curves measured by pump-probe for S1 (43 nm), S2 (78 nm), S3 (121 nm), S4 (165 nm). The excitation wavelength is 800 nm and the probe wavelength is 950 nm.

The slow decay constant (τ₂) is generally considered as the lifetime of electrons in defect levels.³¹ These fitting results suggest that the thicker films (S3, S4) have longer decay lifetimes, indicating the presence of a large density of defect states in the sub-band range, which can trap the excited electrons.

As a result of these correlations, we consider that the defect states play a key role in the above competition mechanism. The presence of defect states creates a defect level in the sub-band range; it can even overlap with the band tail due to the higher density of defect states. That is, the higher the density of defect states, the narrower the band gap will be, which is in agreement with the results in Fig. 5. The low density of defect states in the thinner films (S1, S2) make it easier for electrons to saturate the defect level in the OA Z-scan experiment, but it is more difficult in the thicker films (S3, S4). The density of defect states is the main reason causing the competition mechanism, thereby making the tuning behavior of nonlinear optical absorption.

The usual explanation for the transition is based on TPA-induced ESA with a four-level model.¹⁰ If the switching from SA to RSA was caused by TPA-induced ESA, the electron will be first transferred to conduction band (3.12 eV) and then to a higher energy state (4.68 eV) while absorbing incident photon energy, but both these energy levels are larger than the band gap of In₂Te₃ thin films, which does not meet the requirement of defect-induced absorption. Furthermore, TPA-induced ESA is in fact a kind of fifth order nonlinearity, which can hardly occur in our incident power range (60 GW cm⁻²). Thus we think ESA in a three-level energy model is possible and better explains our observed tuning behaviour in the Z-scan experiment.

Fitting calculation

From the practical point of view, the amorphous In₂Te₃ thin films are in metastable states and they will be transformed to other stable states once the ambient parameters are broken. Now we will focus on the NOA coefficient of crystallized In₂Te₃ thin films at 800 nm wavelength. The contributions of ESA and multi-photon absorption are difficult to distinguish because the absorption coefficient obtained from the Z-scan fitting result is the overall nonlinear absorption effect. Hence, the reported fitting experimental results usually give an effective absorption coefficient; herein, we also defined the effective absorption coefficient β_eff in following formula. The raw data are fitted using typical OA Z-scan theory proposed by Sheik-Behae. The nonlinear coefficient β_eff, defined as α = α₀ + β_effI, can be calculated by fitting the following equation.²⁰where T_OA is the normalized transmittance of the OA Z-scan measurement, z is the longitudinal displacement of the samples, z₀ is the Rayleigh length, is the effective thickness of samples, α₀ is the linear absorption coefficient at 800 nm wavelength, L is the physical thickness of the thin films, and I₀ is the incident laser intensity at the focal point. Typically, only the first few terms of the polynomial are needed for numerical calculation if the series converges.

The Z-scan curves will result in a valley for β_eff > 0, indicative of RSA; a peak for β_eff < 0 in the case of SA. Therefore, both sign and magnitude of β can be evaluated from the fitting curves. In Fig. 6(b), the β_eff values of S1, S2, S3 and S4 at the wavelength of 800 nm are calculated to be as large as −805.6 cm GW⁻¹, −616.2 cm GW⁻¹, 130.6 cm GW⁻¹ and 237.6 cm GW⁻¹, respectively. These values are much larger than those of most semiconductor thin films, such as ZnO,³² InSe,³³ CdO nanomorphotypes and ferroelectric thin films such as Bi_3.15Nd_0.85Ti₃O₁₂ (ref. 8 and 30) at the 800 nm femtosecond pulse excitation and also larger than some organic materials such as CuPc-doped PMMA and graphene oxide metal porphyrin films, but the damage threshold (the largest value of laser pulse intensities that they can sustain) is smaller than these organic materials. It was found that the damage value was observed up to about 500 GW cm⁻² for In₂Te₃ thin films, where an abrupt transmission jump occurs and prominent surface defects can also been observed. Concisely, the large NOA response of In₂Te₃ thin films indicates exhilarating potential applications in ultrafast nonlinear limiting devices. The combination of SA and RSA, merely by thickness regulation through the same material, will become the right candidate in future applications of optical diodes and ultrafast switches.³⁴

Conclusion

In summary, In₂Te₃ thin films of various thicknesses have been successfully prepared using an RF magnetron sputtering method. Basic characterization, including morphology, structure and linear absorption was investigated and the results show that there existed defect states in In₂Te₃ thin films. NOA properties were studied in detail through the OA Z-scan technique using ultrafast pulses at the wavelengths of 800 nm and 400 nm, and the results demonstrated the large NOA response characteristics and tunable NOA behavior in In₂Te₃ films. The large NOA is attributed to the existence of defect states, and tuning behavior originates from the result of an absorption competition mechanism. With the shrinking of band gaps, the competition mechanism is transferred from ground states dominant transition to that of excited states, which is understood by the experimental results. Pump-probe measurement results show the density of defect states plays a critical role in the competition mechanism. The fitting results reveal large NOA coefficients in In₂Te₃ films. Such superior NOA properties render this a promising material for ultrafast nonlinear optical devices and multifunctional nanophotonic components in the future.

Acknowledgements

The authors would like to express their sincere thanks for the financial support from funding under Grant Nos. 60578047, 2009CB929201, 2012CB934303, 13ZR1402600, 06DJ14007, and 2011ZX02402. The authors thank Prof. L. Y. Chen and Prof. M. Xu for effective backup.

References

1. Q. Zheng, H. Zhu, S. Chen, C. Tang, E. Ma and X. Chen, Nat. Photonics, 2013, 7, 234–239.
2. Y. W. Song, S. Yamaashita and S. Maruyama, Appl. Phys. Lett., 2008, 92, 021115.
3. S. R. Mishra, H. S. Rawat, S. C. Mehendale, K. C. Rustagi, A. K. Sood, R. Bandyopadhyay, A. Govindaraj and C. N. R. Rao, Chem. Phys. Lett., 2000, 317, 510–514.
4. J. Wei and H. Yan, Opt. Lett., 2014, 22, 6387–6390.
5. B. Anand, R. Podila, K. Lingam, S. R. Krishnan, S. S. S. Sai, R. Philip and A. M. Rao, Nano Lett., 2013, 13, 5771–5776.
6. D. A. Wheeler and J. Z. Zhang, Adv. Mater., 2013, 25, 2878–2896.
7. M. Rumi and J. W. Perry, Adv. Opt. Photonics, 2010, 2, 451–518.
8. M. G. Kuzyk, J. Mater. Chem., 2009, 19, 7444–7465.
9. Y. B. Band, D. J. Harter and R. Bavli, Chem. Phys. Lett., 1986, 126, 280–284.
10. S. Roy and C. Yadav, Appl. Phys. Lett., 2013, 103, 241113.
11. F. Li, P. Lu, H. Long, G. Yang, Y. Li and Q. Zheng, Opt. Express, 2008, 16, 14571–14581.
12. C. Yadav and S. Roy, J. Comput. Electron., 2015, 14, 209.
13. Q. Ouyang, H. Yu, K. Zhang and Y. Chen, J. Mater. Chem. C, 2014, 2, 6319–6325.
14. Y. Lee, Y. Yan, L. Polavarapu and Q. Xu, Appl. Phys. Lett., 2009, 95, 023105.
15. R. R. Desai, D. Lakshminarayana, P. B. Patel and C. J. Panchal, Sens. Actuators, B, 2005, 107, 523–527.
16. G. Tai, C. Miao, Y. Wang, Y. Bai, H. Zhang and W. Guo, Nanoscale Res. Lett., 2011, 6, 329.
17. M. Emzianea, J. C. Bernède, J. Ouerfelli, H. Essaidi and A. Barreau, Mater. Chem. Phys., 1999, 61, 229–236.
18. R. R. Desai, D. Lakshminarayana, P. B. Patel and C. J. Panchal, J. Mater. Sci., 2006, 41, 2019–2023.
19. C. Wang, X. Bu, N. Zheng and P. Feng, Angew. Chem., 2002, 41, 1959–1961.
20. M. Sheik-Bahae, A. A. Said, T.-H. Wei, D. J. Hagan and E. W. V. Stryland, IEEE J. Quantum Electron., 1990, 26, 760–769.
21. R. A. Ganeev, A. I. Ryasnyansky, N. Ishizawa, M. Baba, M. Suzukia, M. Turu, S. Sakakibara and H. Kuroda, Opt. Commun., 2004, 231, 431–436.
22. A. Singh, S. Kumar, R. Das and P. K. Sahoo, RSC Adv., 2015, 5, 88767–88772.
23. R. R. Desai, D. Lakshminarayana, P. B. Patel and C. J. Panchal, Mater. Chem. Phys., 2005, 94, 308–314.
24. P. Zhou, G. J. You, J. Li, S. Y. Wang, S. X. Qian and L. Y. Chen, Opt. Express, 2005, 13, 1508–1514.
25. J. Tauc, R. Grigorovici and A. Vancu, Phys. Status Solidi, 1966, 15, 627–637.
26. J. P. Xu, R. J. Zhang, Y. Zhang, Z. Y. Wang, L. Chen, Q. H. Huang, H. L. Lu, S. Y. Wang, Y. X. Zheng and L. Y. Chen, Phys. Chem. Chem. Phys., 2016, 18, 3316–3321.
27. K. F. Mak, C. Lee, J. Home, J. Shan and T. Heinz, Phys. Rev. Lett., 2010, 105, 474–479.
28. S. Cheng, S. J. Wei, X. Y. Yi, J. Wang, C. C. Liu, J. Li and T. Y. Yang, J. Phys. D: Appl. Phys., 2015, 48, 475108.
29. J. Liu and J. S. Wei, J. Appl. Phys., 2009, 106, 083112.
30. S. Li, X. L. Zhong, G. H. Cheng, X. Liu, Y. Zhang, J. B. Wang, H. J. Song, C. B. Tan and B. Li, Appl. Phys. Lett., 2015, 106, 142904.
31. P. Dugar, M. Kumar, T. C. Shibin Krishna, N. Aggarwal and G. Gupta, RSC Adv., 2015, 5, 83969–83975.
32. R. U. Haskar, B. Karthikeyan, P. Sreekanth and R. Philip, RSC Adv., 2015, 5, 13590–13597.
33. M. Yüksek, U. Kürüm, H. G. Yaglioglu, A. Elmali and A. Ateş, J. Appl. Phys., 2010, 107, 033115.
34. S. Roy and C. Yadav, Opt. Commun., 2011, 284, 4435.

---