Effect of dopant concentration on the spectroscopic properties in an In³⁺ doped (0, 1, 2 and 4 mol%) Yb:Tm:LiNbO₃ crystal

Abstract

A series of Yb:Tm:LiNbO₃ crystals with x mol% In³⁺ ions (x = 0, 1, 2 and 4 mol%) were grown by a conventional Czochralski technique for the first time. The In, Yb and Tm contents in the crystals were measured by an inductively coupled plasma atomic emission spectrometer (ICP-AES). The results revealed that the segregation coefficients of Yb³⁺ and Tm³⁺ ions decreased with increase of the In³⁺ doping concentration. Also the segregation coefficient of In³⁺ ions increases with the increasing In³⁺ doping concentration in the melts, and it is close to one. The IR transmission spectra and UV-vis spectra were measured to analyze the defect structure of the crystals.

---

1. Introduction

Tm³⁺-based lasers emitting in the wavelength region of 2 μm have attracted much attention in recent years such as in atmospheric, telecommunications, medicinal, and space application fields.^1–3 The line width of Tm³⁺ ions (associated with ³F₄ → ³H₆ transitions) corresponds to the wavelength range of 2 μm, which is similar to Ho³⁺ ions (associated with ⁵I₇ → ⁵I₈ transitions), but the absorption range is wide and the absorption cross section is large. The LiNbO₃ crystal is an excellent material exhibiting desirable characteristics such as electro-optic, acousto-optic, piezo-electric and nonlinear optical performances.^4,5 From a materials point of view, several RE-doped LiNbO₃ have been under study in the form of solid state laser devices and planar waveguide lasers. Yb³⁺ ion co-doped crystals have received considerable attention due to its high absorption cross-section in the near-infrared, suitable for convenient excitation with commercially available optical sources. Recently, many efforts have been devoted to the upconversion luminescence property in RE-doped LiNbO₃ crystal. The basic optical properties of Tm:LiNbO₃ and Yb:LiNbO₃ crystals have been already investigated^6,7 and are considered to be excellent upconversion media. Unfortunately, to our best knowledge, the properties of Yb:Tm:LiNbO₃ crystal is still to be scrutinized. In recent years, 2 μm lasers have attracted a large amount of interests due to their wide tenability and high efficiency, which have increasingly important application in remote sensing, lidar and material processing, etc. It is believed that the study of Yb:Tm:LiNbO₃ crystals will be attractive researches.

The light induced refractive index change, called photorefractive effect, limits the performance of Yb:Tm:LiNbO₃ crystals. Until now, the comprehensive studies on the suppression of the optical damage caused by the high laser intensity in LiNbO₃ crystal have been investigated. As an effective solution to this problem, it has been found that the significant resistance to optical damage in LiNbO₃ can be improved by doping with certain divalent, trivalent or tetravalent cations (such as Mg²⁺, Zn²⁺, In³⁺, Sc³⁺, Hf⁴⁺ and Zr⁴⁺) above the threshold concentration, see ref. 8–13.

In this letter, In:Yb:Tm:LiNbO₃ crystals with different concentrations of In₂O₃ (0, 1, 2 and 4 mol%) were successfully grown by the Czochralski method. The defect structures were analyzed via IR transmission spectrum and the UV-vis absorption spectra. Finally, the mechanism of the relationship between the spectroscopic properties and the concentration of In₂O₃ was thoroughly discussed.

2. Experiments

In this experiment, congruent LiNbO₃ crystals codoped with 0.5 mol% Yb₂O₃, 0.5 mol% Tm₂O₃, and various concentration of In₂O₃ were grown. The composition of melt was chosen as Li/Nb = 48.6/51.4. The In₂O₃ concentrations were 0, 1, 2 and 4 mol%, respectively, which were labeled as 1#, 2#, 3# and 4#, as are shown in Table 1. All raw materials precisely weighed and thoroughly mixed for 24 h, heated up to 750 °C for 2 h to remove CO₂, and then further sintered at 1050 °C for 2 h. The crystals were grown in open atmosphere by traditional Czochralski method along ferroelectric z-axis with the rotation rate of 15–20 rpm and pulling speed of 1–1.5 mm h⁻¹. After growth, the crystals were cooled down to room temperature at a rate of 60 °C h⁻¹. The crystals were polarized at 1200 °C with a current density of 5 mA cm⁻². The wafers with diameter of 10 × 20 × 2 mm (x × z × y) were cut from the middle part along y-face and polished to optical grade both of sides for optical tests. Fig. 1 showed the wafers of four In:Yb:Tm:LiNbO₃ samples in experiment.

Table 1 Composition of In:Yb:Tm:LiNbO₃ crystal and the size of crystal

Sample	1#	2#	3#	4#
In^3+ in the crystal (mol%)	—	0.762	1.654	3.956
Keff of In^3+	—	0.762	0.827	0.989
Yb^3+ in the crystal (mol%)	0.848	0.793	0.754	0.712
Keff of Yb^3+	0.848	0.793	0.754	0.712
Tm^3+ in the crystal (mol%)	0.877	0.849	0.831	0.816
Keff of Tm^3+	0.877	0.849	0.831	0.816
Crystal size (mm^2)	Φ25 × 28	Φ25 × 38	Φ25 × 35	Φ25 × 25
Axial temperature gradient (°C cm^−1)	25–30	30–45	30–35	30–35
Radial temperature gradient (°C cm^−1)	4–6	4–6	3–5	3–5

---

Fig. 1 Wafers of four In:Yb:Tm:LiNbO₃ crystals in experiment.

We analyzed the concentration of In³⁺, Yb³⁺ and Tm³⁺ in the crystals with an inductively coupled plasma atomic emission spectrometer (ICP-AES, Optima 5300DV). The X-ray powder diffraction patterns were recorded using a SHMADZU XRD-6000 diffractometer with monochromatic Cu Kα radiation. The OH⁻ absorption spectra of In:Yb:Tm:LiNbO₃ crystals were recorded by an Avator-360 FT-IR spectrometer at room temperature. The scanning was carried out between 400 and 4000 cm⁻¹ with the scanning step of 1 cm⁻¹. In our later discussion, 3400 to 3600 cm⁻¹ was chosen as the effective range. The ultraviolet-visible (UV-vis) absorption spectra of the samples were measured with a CARY UV-visible spectrometer, with the incident light transmitting along the y axis at room temperature. The measurement range was from 300 to 1100 nm.

3. Results and discussion

Preparation of standards solution: 99.99% purity In₂O₃, Yb₂O₃ and Tm₂O₃ were put in polytetrafluoroethylene (PTFE) beaker, plus nitric acid and hydrofluoric acid solution drops, and then dissolved by heating to make-up 40 g L⁻¹ stock solution and saved in PTFE bottle. From each top, middle and lower parts of the crystal, cut out a small crystal specimens and placed in an agate mortar grind into fine powder, with the balance millionth 0.05 g powder sample is weighed out. Powder was placed in a PTFE beaker, added hydrofluoric acid and nitric acid solution each 5 mL. After preheated for 25 minutes on a heated plate, using fully automated microwave digestion oven to microwave sample digestion. Then, the sample was heated and titrated with perchloric acid to eliminate hydrofluoric acid and protect the testing equipment from corrosion. At the same time, the digestive process was finished. After cooling to room temperature, the sample solution was mixed with 5% of nitric acid and played in the 50 ml flask for testing, the volume of the solution was set up by addition of 5% of nitric acid. Fig. 1 showed the photos of the In:Yb:Tm:LiNbO₃ crystals samples.

The concentrations of In³⁺, Yb³⁺ and Tm³⁺ ions in the In:Yb: Tm:LiNbO₃ crystals and effective distribution coefficient (K_eff) of ions are listed in Table 1. The effective distribution coefficient of In³⁺, Yb³⁺ and Tm³⁺ ions are calculated by comparing the measured In³⁺, Yb³⁺ and Tm³⁺ ions concentration in the crystal and melt. The dependence of the segregation coefficients of In³⁺, Yb³⁺ and Tm³⁺ ions on the In₂O₃ concentration in the melts is shown in Fig. 2. It is clearly seen that the effective distribution coefficient of In³⁺ ion increases with the increasing In₂O₃ concentration in the melts, with the maximum close to one, indicating that it distributes uniformly in crystal. It can be further seen from Fig. 2 that the effective distribution coefficient of Yb³⁺ and Tm³⁺ ions decreases as the doping concentration of In³⁺ increases in the melts. When In³⁺ ion entered in Yb:Tm:LiNbO₃ crystal, it could readily occupy intrinsic defect (Nb_Li⁴⁺) sites and the Nb_Li⁴⁺ concentration was sharply reduced, which suppressed the entrance of Yb³⁺ and Tm³⁺ ions into crystal. Thus, the effective distribution coefficient of ions Yb³⁺ and Tm³⁺ was continuously reduced.

Fig. 2 The dependence of the segregation coefficient of In, Yb, and Tm ions on the In₂O₃ concentration in the melts.

The XRD experimental results are shown in Fig. 3. It can be seen that the dope of In³⁺, Yb³⁺ and Tm³⁺ ions did not create new diffraction peak. This verified that after adding the In³⁺, Yb³⁺ and Tm³⁺ ions to crystal lattice of LiNbO₃, the structure did not change dramatically.¹⁴ Based on the ion radius approximation principle in the crystal field theory, the doped ion entered crystal lattice by replacing the Li⁺ and Nb⁵⁺ ion instead of occupied the slot among crystal lattice. Hence to obtain the formulated In:Yb:Tm:LiNbO₃ crystals solid solution as a single phase.

Fig. 3 X-ray powder diffraction patterns of In:Yb:Tm:LiNbO₃ crystals.

In course of growth, H₂O vapors can be introduced into the crystals through formation of O–H bonds. The OH⁻ absorption spectra measurement results of In:Yb:Tm:LiNbO₃ crystals are represented in Fig. 4. Fig. 4 summarizes the OH⁻ absorption peak of samples 1#, 2#, 3# and 4# are located at 3482, 3482, 3507 and 3508 cm⁻¹, respectively. It can be found that the spectra belonging to In³⁺ of different concentration form different groups in In:Yb:Tm:LiNbO₃ crystals. According to Li-site vacancy model, in pure congruent LiNbO₃ crystal, the ratio of [Li]/[Nb] is 0.946. There always exist intrinsic lattice defects, such as anti-site Nb (Nb_Li⁴⁺) and Li vacancy (V_Li⁻) defects. The electropositive H⁺ ions easily attracted by the electronegative V_Li⁻ and formed V_Li⁻–OH⁻ complexes, whose vibrational absorption peak is located at 3482 cm⁻¹.¹⁵ In Yb:Tm:LiNbO₃ crystals (i.e. for 1# sample), Yb³⁺ and Tm³⁺ replace Nb_Li⁴⁺, forming Yb_Li²⁺ and Tm_Li²⁺. When the concentration of In³⁺ ions is 1 mol% in the melt, below its threshold value (i.e. for 2# sample), Yb³⁺ and Tm³⁺ ions substitute for Nb_Li⁴⁺ defects and exist in the form of Yb_Li²⁺ and Tm_Li²⁺ in LiNbO₃, while these defects repel the H⁺ and affect the V_Li⁻–OH⁻ complexes only as ions around them and the peak position is at about 3482 cm⁻¹. When the concentration of In³⁺ ions exceeded the threshold (i.e. for 3# and 4# samples), all Nb_Li⁴⁺ defects disappear and In_Nb²⁻ defect structures appears as the In³⁺ ions enter the normal Nb sites. Meanwhile, the amount of Yb³⁺ and Tm³⁺ located at Nb sites increases, namely, the amount of Yb_Nb²⁻ and Tm_Nb²⁻ increases. The In_Nb²⁻ in the In_Nb–H–O defect structure has intensive attraction to H⁺, more energy is required for O–H vibration, and the corresponding absorption is dominant at about 3507 cm⁻¹. According to the analysis above, we can conclude that the In doping threshold concentration is about 1.5–2.0 mol% as previously reported in.^16–18.

Fig. 4 The IR absorption band related to the OH⁻ vibration in In:Yb:Tm:LiNbO₃ crystals.

To confirm our significant conclusion, we further studied the UV-visible absorption spectra of four samples cut from the top, center and bottom of the grown crystals. Fig. 5 shows the UV-visible absorption spectra of the In:Yb:Tm:LiNbO₃ crystals with different concentration of In³⁺. The absorption edge of LiNbO₃ crystals is used to determine the composition.¹⁹ The absorption edge at the level of a = 15 cm⁻¹ or a = 20 cm⁻¹ is usually used to quantify the defects in a LiNbO₃ crystal.²⁰ In our experiment, the referred UV-vis optical absorption coefficient for determination of the position defined as 20 cm⁻¹. It can be seen from Fig. 5 that the UV-vis absorption edges of the 1#, 2#, 3# and 4# samples are located at 353, 346, 335 and 379 nm, respectively. The absorption peak of In:Yb:Tm:LiNbO₃ crystals is first violet-shifted with the increasing In₂O₃ doping concentration in the melt, then red-shifted when the doping concentration of In³⁺ is 4 mol%. The shift of absorption edge (4# sample) is up to 26 nm with respect to position of the Yb:Tm:LiNbO₃crystal. It is a remarkable fact that when In³⁺ doping concentration reaches 4.0 mol%, namely 4# sample, the absorption edge of shifts to the red substantially relative to 3#. There is some absorption band located at 300–1050 nm, which is attributed to absorption of the Tm³⁺and Yb³⁺ions. The three absorption bands of Tm³⁺ ions centered at 465 nm, 692 and 798 nm are assigned to the ³H₆ → ¹G₄, ³H₆ → ³F_2,3 and ³H₆ → ³H₄ transitions.²¹ And 980 nm absorption peak correspond to transition from the ²F_7/2 ground state of Yb³⁺ ion to the excited state ²F_5/2.

Fig. 5 The UV-vis absorption spectra of the In:Yb:Tm:LiNbO₃ crystals.

It's common knowledge that the shift of absorption edge can be explained by the polarization ability of the doped ions.^22,23 The valence electron transition energy, which is from the 2p orbits of O²⁻ to the 4d orbits of Nb⁵⁺, determines the fundamental absorption edge of the LiNbO₃ crystal.²³ It can be concluded from the previous researches that the violet shift of the absorption edge is attributed to the increasing in the ability of adjacent dopant ions to polarize an O²⁻ ion shortens the band gap of the valence electron transition. Otherwise, the absorption edge of sample shifts toward the infrared band. In fact, the polarization ability of Nb⁵⁺ is smaller than that of In³⁺, Yb³⁺, Tm³⁺ and Li⁺. In Yb:Tm:LiNbO₃ crystal, Yb³⁺ ions and Tm³⁺ ions replace Nb_Li⁴⁺ and form Yb_Li²⁺ and Tm_Li²⁺. When the In³⁺ enter Yb:Tm:LiNbO₃ crystals, because of the polarization ability of In³⁺ ions is much smaller than that of Nb_Li⁴⁺, the absorption edge of In:Yb:Tm:LiNbO₃ crystal (2# sample) shifts to the ultraviolet band compared with that of Yb:Tm:LiNbO₃ crystal (1# sample). When the concentration of In³⁺ ions in the melt exceeded the threshold value (3# sample), Nb_Li⁴⁺ defects almost disappeared, the In³⁺ ions began to enter the normal Li and a little bit of Nb sites. As the polarization ability of In³⁺ was lower than that of Li⁺, which caused the deformation of O²⁻ ions electron clouds, the width of the forbidden band increases and absorption edge continuous shifts to the violet. When the In₂O₃ doping concentration in the melts is up to 4.0 mol% (4# sample), part of In³⁺, Yb³⁺ and Tm³⁺ ions begin occupies Nb sites form In_Nb²⁻, Yb_Nb²⁻ and Tm_Nb²⁻. As the polarization ability of In³⁺ ion is smaller than that of Nb⁵⁺, red shift of the absorption edge in 4# relative to 1#. This discussion of dopant occupancy is consistent with the analysis result of experimental shown in OH⁻ absorption spectra.

4. Conclusions

In conclusion, In:Yb:Tm:LiNbO₃ crystals with different concentrations of In₂O₃ (0, 1, 2 and 4 mol%) were successfully grown by the Czochralski method. The inductively coupled plasma optical emission spectrometry, IR transmission spectrum and the UV-vis absorption spectra were measured and discussed in terms of the spectroscopic properties in In³⁺ doped (0, 1, 2 and 4 mol%) Yb:Tm:LiNbO₃ crystal. The effective distribution coefficient of In ion increases with the increasing In₂O₃ concentration in the melts, and it is close to one, indicating that In³⁺ distributes uniformly in crystal. The UV-vis absorption spectra showed that the absorption edges of In:Yb:Tm:LiNbO₃ crystals is first violet-shifted with the increasing In₂O₃ doping concentration in the melt, then red-shifted when the doping concentration of In³⁺ is 4 mol%. It was suggested that for low concentration of In³⁺ ion, the all dopant ions first occupy the Nb anti-sites or Li vacancies and when the concentration of In³⁺ was reaches 4.0 mol% they may replace normal Nb sites.

Acknowledgements

This work is supported by National Natural Science Foundation of China (no. 51301055), Provincial University Key Lab Open Foundation for Materials Research and Application in Harbin University of Science and Technology (no. 2013) and the Opening Project of State Key laboratory of Crystal Material (no. KF1409).

References

1. A. Bensalah, K. Shimamura and V. Sudesh, et al., J. Cryst. Growth, 2001, 223, 539–544.
2. R. A. McFarlane, J. Opt. Soc. Am. B, 1994, 11, 871–880.
3. C. Li, X. Zhang and C. Wang, et al., Chin. J. Inorg. Chem., 2011, 27, 6–10.
4. D. Janner, D. Tulli, M. García-Granda, M. Belmonte and V. Pruneri, Laser Photonics Rev., 2009, 3, 301–313.
5. T. Y. Fan, A. Cordova-Plaza, M. J. F. Digonnet, R. L. Byer and H. J. Shaw, J. Opt. Soc. Am. B, 1986, 3, 140–148.
6. M. Jelínek, J. Oswald, T. Kocourek, K. Rubešová, P. Nekvindová, D. Chvostová, A. Dejneka, V. Železný, V. Studnička and K. Jurek, Laser Phys., 2013, 23, 105819.
7. V. T. Pham, S. K. Lee, M. T. Trinh, K. S. Lim, D. S. Hamilton and K. Polgar, Opt. Commun., 2005, 248, 89–96.
8. T. R. Volk and N. M. Rubinina, Ferroelectr., Lett. Sect., 1992, 14, 37–43.
9. X. Zhen, Q. Li, L. Wang and Y. Xu, J. Cryst. Growth, 2005, 284, 270–274.
10. S. Li, S. Liu, Y. Kong, J. Xu and G. Zhang, Appl. Phys. Lett., 2006, 89, 101126–101134.
11. Z. Xu and Y. Xu, Mater. Lett., 2007, 61, 3243–3246.
12. L. Dai, Y. Su, S. Wu, J. Guo, C. Xu and Y. Xu, Opt. Commun., 2011, 284, 1721–1725.
13. L. Sun, F. Guo, Q. Lv, H. Yu, H. Li and W. Cai, Cryst. Res. Technol., 2007, 42, 1117–1122.
14. Y. Fan, A. Cordova-Plaza and M. J. F. Digonnet, J. Opt. Soc. Am. B, 1986, 3, 140.
15. O. F. Schirmer, O. Thiemann and M. Wöhlecke, J. Phys. Chem. Solids, 1991, 52, 185–200.
16. T. Volk, N. Rubinina and M. Wöhlecke, J. Opt. Soc. Am. B, 1994, 11, 1681–1687.
17. N. Y. Kamber, J. Xu, S. M. Mikha, G. Zhang and G. Zhang, Opt. Commun., 2000, 176, 91–96.
18. H. Qiao, J. Xu, Q. Wu, X. Yu, Q. Sun, G. Zhang, X. Zhang and T. Volk, Opt. Mater., 2003, 23, 269–277.
19. I. Földvári, K. Polgár and R. Voszka, Cryst. Res. Technol., 1984, 19, 1659–1661.
20. K. Polgár, Á. Péter and L. Kovács, J. Cryst. Growth, 1997, 177, 211–216.
21. C. Sun, F. Yang, T. Cao, Z. You, Y. Wang, J. Li, Z. Zhu and C. Tu, J. Alloys Compd., 2011, 509, 6987–6993.
22. L. Dai, C. Xu, Z. Qian, J. Li, D. Li and Y. Xu, J. Lumin., 2013, 134, 255–259.
23. W. Wang, X. Chen, D. Ni, D. Zhang and X. Wu, J. Alloys Compd., 2005, 402, 224–226.

---