We report on the light scattering phenomenon in annealed multidomain LiNbO3:Fe:Hf crystals. The scattering sources are found to be some fog-like “defects”, which cause the polarization-dependent scattering of the light, and can be removed completely by the illumination of visible light. Based on these results and the etch patterns, these “defects” are suggested to be refractive index fluctuations induced by the space charges accumulated at the boundary of opposite microdomains. The influence of quick heating-up on the “defects” is also studied and the results firmly support our suggestion about the nature of the “defects”. At last, the temporal curves of the transmitted intensity during the light scattering are explained. The mechanism for the opposite microdomain formation is also explained from the view of crystal growth.
©2010 Optical Society of America
In the past several decades, Fe-doped lithium niobate (LiNbO3:Fe) crystals are extensively studied due to their many photorefractive applications . Codoping with damage-resistant elements, such as Mg, Zn and In, is an efficient way to tailor the photorefractive properties of LiNbO3:Fe crystals [2–5]. Recently, Hf element was codoped in LiNbO3:Fe crystals as a new damage-resistant dopant, and Fe ions were found to always occupy Li sites in spite of Hf incorporation . Due to this specialty of Hf doping, LiNbO3:Fe:Hf crystals exhibit unusual light scattering behavior as compared with other codoped ones .
Compared with Mg, Zn and In, tetravalent Hf not only has a different influence on the optical properties of LiNbO3:Fe crystals, but also affects their domain behavior in a different way. Previous studies have shown that high Mg doping can convert the multidomain nature of LiNbO3:Fe crystals to a complete singledomain nature [8, 9]. It was suggested that this change in domain structure is due to the reduction of the intrinsic defects in LiNbO3:Fe:Mg crystals . However, for Hf element, Bermúdez et al. reported a dissimilar behavior. They showed that 1-mol%-Hf-doped LiNbO3 crystals grown from a congruent melt containing K2O still exhibit a multidomain nature . Moreover, they studied the PPLN domain formation during off-center CZ crystal growth, and found that the domain structure has a close connection with the local Hf concentration modulation . These results imply that Hf doping may disturb the stability of a singledomain structure, which is in contrast with the role of Mg doping on domain formation.
In this paper, we report on the light scattering phenomena in multidomain LiNbO3:Fe:Hf crystals. These phenomena are found closely connected with the photorefractive effect and especially with the opposite microdomains formed in the crystals.
2. Experimental procedures and results
Samples used in this study were LiNbO3 crystals doped with 0.03 mol% Fe and 5.0 mol% Hf. The damage-resistant doping concentration of the samples exceeds the so-called threshold, based on the results given by Razzari et al . The highly doped LiNbO3 crystal was grown along z-axis from the congruent melt by the Czochralski technique. Neither poling operation in CZ system nor post-growth poling step was employed. The as-grown crystal was cut to rectangular-shaped y-oriented plates and then polished to optical grade.
By using the selective decoration method , the domain structure of samples was found to be a multidomain one, which is consistent with the results given by Bermúdez et al . It should be noticed that the as-grown samples are clear, color-uniform and completely transparent. However, after the heat treatment at 200°C for 0.5h, their optical status exhibits an obvious change. The samples look no longer clear and some fog-like “defects” are present in the bulk. These defects can be observed directly by naked eyes and distribute uniformly in the crystals. In order to investigate the nature of fog-like “defects”, the conventional light scattering experiments were carried out on these samples.
Experimental arrangement of light scattering experiment is shown in Fig. 1 . Expended laser beams, with the wavelength of 1064, 780, 633 and 514nm, were used as pump light (denoted as “P” in Fig. 1). The beam impinged on the sample along y axis and its spot size on the sample was approximately 3mm. The pump light intensity was varied using a neutral density filter and a screen was used to block the scattered light. In addition, a reference light (denoted as “R” in Fig. 1) was taken out from the incident beam for monitoring the drifts caused by the laser power fluctuation. During the experiment, the transmitted pump light was collected by a detector and the temporal curve of transmitted intensity was recorded by a PC.
The experiments show that the fog-like “defects” induce obvious scattering of the pump light for all wavelengths and make the crystals opaque. Attention should be paid to two main results. Firstly, large difference exists between the intensity of the scattered light for e- and o- polarization: e-polarized light is scattered by the “defects” far more seriously than o-polarized one. Secondly, the “defects” can be removed by the continuous illumination of visible light (633nm and 514nm), and the illumination time for the complete removal of the “defects” decreases with the increase of the pump intensity. Figure 2 shows the temporal curves of transmitted intensity for 514nm light with different pump intensities (I0), where (a) and (b) correspond to o and e-polarization, respectively. Through dividing the initial transmitted intensity by the pump intensity I0, we can get the initial transmittance of the sample, which are 50% and 6% for o and e-polarization, respectively. This transmittance contrast clearly shows the polarization dependence of the light scattering. Furthermore, an obvious light-induced transparence can be found in the temporal curves for both polarizations, which is the consequence of the “defects” removal by the visible light illumination. In fact, we find in the experiments that, the 514nm light is more effective than the 633nm on removing the “defects”.
Fe2+/Fe3+ traps, the well-known photorefractive centers in LiNbO3:Fe crystals, are very sensitive to the green light . The similar sensitivity of the “defects” to the green light indicates their connection with the photorefractive effect. In addition, it is known that photorefractive effect for e-polarization is much stronger than that for o-polarization . And the light scattering in our work shows the same behavior: e-polarized light is scattered more seriously than o-polarized one. Thus, we believe that the “defects” come into being through the photorefractive effect. In details, they have a nature of refractive index fluctuation (Δn) and are induced by the space-charge field at micrometer scale.
It has been pointed out in the introduction that, Hf doping may disturb the stability of a singledomain structure, and also, our as-grown crystal is found to be multidomain. Thus, it is highly possible that a large amount of microdomains exists in the crystal and is responsible for the source of space charges. In order to prove this argument, the sample with fog-like “defects” was cut into several c plates, polished and etched in HF acid for domain observation. Figure 3 gives the etch patterns of the c-cut surface of the sample. It is obvious that inside the crystal indeed lots of microdomains exist with the size ranging from submicron to tens micron. Moreover, these microdomains should be opposite domains, because the similar microdomain distribution was found on the y-cut surface.
Based on the above results, the mechanism for the formation and removal of the “defects” can be given as follow (see Fig. 4 ): the positive bound charges accumulate at “head to head” domain boundary and the negative ones get together at “tail to tail” boundary, which induces localized strong space charge field and in turn causes lots of micrometric refractive index fluctuations (Δn). When pump light propagates in the crystal, it is scattered by the micrometric refractive index fluctuations, what makes the crystal opaque. As the visible light illumination is present, however, electrons will be excited from Fe2+ into the conduction band and compensate rapidly the space charges. This process makes the space charge field to disappear and, consequently, the field-induced refractive index fluctuations disappear as well. This explains why the fog-like “defects” can be removed by the illumination.
In fact, the accumulation process of the bound charges is the result of a competition between the pyroelectric effect and the compensation effect of mobile H+ in the crystal. It is well known that the pyroelectric effect can be considered as the spontaneous polarization change ΔP with a temperature variation ΔT, according to ΔP=pΔT, where p is the pyroelectric coefficient and has a negative value for LiNbO3 crystals . In our case, when the crystal temperature is high enough (e.g. 200°C), the bound charges at the boundary of opposite domains can be totally compensated by the mobile H+. However, when the crystal is cooled down below 80°C, the mobility of H+ decreases a lot  but the polarization magnitude continues to increase, which makes the bound charges no longer compensated. As a result, the bound charges accumulate and the “defects” form. This qualitative scheme of the charge dynamic in the crystal explains why the “defects” occur after the heat treatment. On the other side, if the “defects” have been formed in the crystal, the heating-up, in a reversed way, will make the polarization magnitude to decrease, thus decreasing the space charges accumulated in the fog-like area and making this area transparent. Moreover, if a part of the fog-like area generated after annealing was made clear by visible light illumination, that is by totally compensating the bound charges through Fe2+/Fe3+ centers, then the electrostatic equilibrium will be broken by the polarization magnitude variation in the heating process. As a result, the fog-like “defects” will be present in this area after the heating-up. Figure 5 shows the pictures of the sample before and after the quick heating-up, where the heating process elevates the sample temperature above 80°C. It is obvious that the light-induce-transparent area and the fog-like area exchange their status after the quick heating-up. This experimental result supports firmly our hypothesis about the nature of the “defects”.
From Fig. 2, it is easily seen that the temporal curves for o-polarization and e-polarization are very different in shape. For o-polarization, only the light-induced transparence is present in the curve. But for e-polarization, an obvious intensity fluctuation is superimposed on the initial part of the light-induced transparence curve. We attribute this intensity fluctuation to the typical beam fanning effect in LiNbO3:Fe:Hf crystals, because this effect has been found to influence the transmitted intensity deeply when pump intensity is above 400mW/cm2 . The absence of the intensity fluctuation from the curves for o-polarization is possibly due to the week photorefractive effect for o-polarization. Figure 2(b) shows that the curve fluctuation varies with the pump intensity. In comparison with the case for 450mW/cm2, the curve fluctuation for 750mW/cm2 is larger and occurs earlier. It is because the beam fanning effect becomes stronger and faster with the increase of the pump intensity. However, the fluctuation disappears from the curve for 1130mW/cm2. The reason for this may be that the beam fanning in this pump intensity occurs so fast that it is overwhelmed by the strong light-induced transparence at the beginning of illumination.
As already mentioned in the experimental section, neither poling operation in CZ system nor post-growth poling step was employed in the crystal fabrication procedures. Unpoled LiNbO3:Fe crystals were also investigated for comparison, but no similar kind of light scattering and light-induced transparence phenomenon was found in these crystals. As a matter of fact, unpoled LiNbO3:Fe crystals contain lots of macroscopic multidomains and few opposite microdomains, which is in agreement with the crystal etch patterns in Ref. 9. These results indicate that Hf doping plays a key role in the formation of the opposite microdomains. In fact, the mechanism for opposite microdomain formation can be explained from the view of crystal growth. As already described in the introduction, high Mg doping in LiNbO3 crystal always leads to a complete singledomain nature [8,9], but Callejo et al  demonstrated clearly that, Hf doping is very special as compared with other doping and that, the spatial variation of Hf concentration in the crystal can lead to the change of local domain polarization. In the work of Callejo et al , the spatial variation of Hf concentration was introduced designedly through the off-center growth technique in order to realize the control of domain pattern at micrometer scale. In our case, no off-center growth technique was employed, but the variation of Hf concentration still exists because of the high impurity concentration in the melt and some unavoidable environmental influences (e.g., mechanical vibration, external temperature fluctuation, heating power fluctuation) on the local growth rate at the solid-liquid interface . Apparently, Hf concentration variation in this case is out of control and consequently leads to the random formation of opposite microdomains in the crystal. For getting a single domain crystal, a dc electrical field must be applied to the crystal-melt system to suppress the formation of microdomains, just as done in the precise works of Razzari and Kokanyan et al. [11, 16]. In addition, post-growth poling method was also tried in our work and it was found valid for removing the microdomains in the crystals.
In this paper, we report on the light scattering phenomenon in annealed multidomain LiNbO3:Fe:Hf crystals. The scattering sources are found to be some fog-like “defects” induced by the heat treatment. The light scattering experiments were carried out to investigate these “defects”. The results show that e-polarized light is scattered by the “defects” far more seriously than o-polarized one and that the “defects” can be removed by the continuous illumination of the visible light. Based on these results and the etch patterns of the sample surface, it was suggested that the “defects” have the nature of refractive index fluctuation and they are induced by the space charges accumulated at the boundary of the opposite microdomains. The influence of quick heating-up on the “defects” is also studied. It is observed that the light-induce-transparent area and fog-like area exchange their status after the quick heat-up. This experimental result firmly supports our suggestion about the nature of the “defects”.
This work is partly supported by Hebei Natural Science Foundation (No. F2009000108 and F2007000119), the Tianjin Natural Science Foundation (No. 09JCYBJC02400), and Hebei Education Bureau Project (No.2008113). We would like to thank the referees for their valuable comments.
References and links
1. P. Günter, and J. P. Huignard, Photorefractive Materials and Their Applications, Vols. I and II, (Springer, Heidelberg, 1989).
2. T. Volk, N. Rubinina, and M. Wöhlecke, “Optical-damage-resistant impurities in lithium niobate,” J. Opt. Soc. Am. B 11(9), 1681–1687 (1994). [CrossRef]
3. T. Volk, and M. Wöhlecke, Lithium Niobate: Defects, Photorefraction and Ferroelectric Switching, (Springer, Berlin, 2008).
4. T. Volk, M. Wöhlecke, N. Rubinina, N. V. Razumovski, F. Jermann, C. Fischer, and R. Bower, “LiNbO3 with the damage-resistant impurity indium,” Appl. Phys., A Mater. Sci. Process. 60, 217–225 (1995). [CrossRef]
5. M. Simon, F. Jermann, T. R. Volk, and E. Krätzig, “Influence of zinc doping on the photorefractive properties of lithium niobate,” Phys. Status Solidi 149(2), 723–732 (1995). [CrossRef]
6. S. Li, S. Liu, Y. Kong, J. Xu, and G. Zhang, “Enhanced photorefractive properties of LiNbO3:Fe crystals by HfO2 co-doping,” Appl. Phys. Lett. 89(10), 101126 (2006). [CrossRef]
7. W. Yan, H. Chen, L. Shi, S. Liu, and Y. Kong, “Investigations of the light-induced scattering varied with HfO2 codoping in LiNbO3 Fe crystals,” Appl. Phys. Lett. 90(21), 211108 (2007). [CrossRef]
8. A. M. Prokhorov, and Yu. S. Kuz’minov, Physics and Chemistry of Lithium Niobate, (Bristol, Hilger, 1990)
9. V. Bermúdez, P. S. Dutta, M. D. Serrano, and E. Dieguéz, “The effect of native defects on the domain structures of LiNbO3: Fe- a case study by addition of MgO and K2O to the congruent melt,” J. Phys. Condens. Matter 9(28), 6097–6101 (1997). [CrossRef]
10. D. Callejo, V. Bermúdez, and E. Dieguéz, “Influence of Hf ions in the formation of Periodic Poled Lithium Niobate Structures,” J. Phys. Condens. Matter 13(6), 1337–1342 (2001). [CrossRef]
11. L. Razzari, P. Minzioni, I. Cristiani, V. Degiorgio, and E. P. Kokanyan, “Photorefractivity of Hafnium-doped congruent lithium–niobate crystals,” Appl. Phys. Lett. 86(13), 131914 (2005). [CrossRef]
12. E. Soergel, “Visualization of ferroelectric domains in bulk single crystals,” Appl. Phys. B 81(6), 729–751 (2005). [CrossRef]
13. K. Buse, “Light-induced charge transport processes in photorefractive crystals II: Materials,” Appl. Phys. B 64(4), 391–407 (1997). [CrossRef]
14. R. S. Weis and T. K. Gaylord, “Lithium Niobate: Summary of Physical Properties and Crystal Structure,” Appl. Phys., A Mater. Sci. Process. 37(4), 191–203 (1985). [CrossRef]
15. A. Yariv, S. S. Orlov, and G. A. Rakuljic, “Holographic storage dynamics in lithium niobate: theory and experiment,” J. Opt. Soc. Am. B 13(11), 2513–2523 (1996). [CrossRef]
16. E. P. Kokanyan, L. Razzari, I. Cristiani, V. Degiorgio, and J. B. Gruber, “Reduced photorefraction in hafnium-doped single-domain and periodically poled lithium niobate crystals,” Appl. Phys. Lett. 84(11), 1880–1882 (2004). [CrossRef]