Open Access
Add to BibSonomyAbstract
We study the temperature and pump-intensity dependence of the temporal evolution of photoinduced light scattering in congruent, hafnium-codoped and near-stoichiometric LiNbO3:Fe crystals. A partial self-recovery process is found in the evolution trend of photoinduced light scattering in all crystals at room temperature while it disappears at −190 °C or above 120 °C. Moreover, a “scattering acceleration effect” is found for both Hf-doped and near-stoichiometric crystals. Basing the saturation refractive-index modulation the partial self-recovery process is explained by the slow recombination of light-induced space charges of the noise gratings with high wave vectors. The “scattering acceleration effect” is connected with the activation of the protons located around Fe ions and their impact on the photoconductivity.
© 2016 Optical Society of America
1. Introduction
Propagation of a laser beam in LiNbO3:Fe (LN:Fe) crystal often causes the photoinduced light scattering (PILS) effect, which leads to the significant decrease of main beam transmittance [1]. The PILS phenomenon was studied very early and many experimental results and theoretical models have been reported for PILS [2–11]. Segev et al. showed that the PILS starts from beam coupling between the incident beam and part of the incident radiation scattered by noise at or near the input plane [2]. On one hand, the PILS effect may strongly reduce the signal-to-noise ratio and hinder the application of LN:Fe crystal in holographic data storage [3]. On the other hand, the PILS can be utilized as a powerful tool for analyzing the various characteristics of LN crystals [4–11]. For example, the PILS was utilized by Ellabban et al. as a technique to determine the activation energy for the thermal fixing in LN:Mn [4]. Goulkov et al. considered the PILS as a normal two-wave mixing process taking into account the competition of photovoltaic and diffusion fields and the contribution of “hot electrons” in the space-charge transport [5–7]. They developed a one-shot PILS-based method to measure all parameters describing the holographic behaviors of LN [6]. Through the PILS effect induced by focused laser beam, Kostritskii et al. studied the role of intrinsic defects in the photoinduced charge transport for LN and the influence of steady-state temperature gradient on photorefractive damage [8–10]. Despite the intensive studies of the PILS in LN, one open issue regarding the origin of the self-recovery of main beam transmittance still remains. Zhang et al. performed systematical PILS experiments on LN:Fe:In crystals with different doping levels, and found a partial self-recovery of main beam transmittance during the PILS in moderate In-codoped LN:Fe crystals [11]. The competition between the dc and ac components of refractive index perturbation was proposed to explain this process. However, the physical essence of this competition was not given in their work. Kostritskii et al. [10] found the similar PILS process called as “self-compensation of optical damage” in chemically reduced pure LN crystals. They suggested that the steady-state temperature gradient generated by the inhomogeneous local heating of the illuminated area give rise to the self-recovery of the transmittance. Obviously, more experimental results are required for better understanding this self-recovery process of main beam transmittance during PILS.
Changing the Li composition of LN:Fe or codoping LN:Fe with tetravalent Hf ions may largely modify its photorefractivity [12–14]. Thus, the PILS behaviors in the modified LN:Fe may provide information about the self-recovery process. In addition, the behaviors of the PILS at different temperatures are expected to be more informative. In this paper, we study the temperature and pump-intensity dependence of the temporal evolution of PILS in congruent, Hf-codoped and near-stoichiometric LN:Fe crystals. The self-recovery process will be connected with the slow recombination of photoinduced space charges of the noise gratings with high wave vectors. Moreover, “scattering acceleration effect” will be reported in Hf-codoped and near-stoichiometric LN:Fe crystals.
2. Samples and experiments
Samples used in this work include y-cut congruent LN doped with 0.028 mol% (0.03 wt%) Fe2O3, LN codoped with 5mol% HfO2, and near-stoichiometric LN ([Li]/[Nb] + [Li] = 49.3%). All samples are about 1mm-thick and denoted as CLN:Fe, CLN:Fe:Hf and SLN:Fe, respectively. The narrow Raman peaks of SLN:Fe (See Fig. 1(a)) indicate its lattice is relatively close to the perfection. This point could also be reflected from the OH- vibration bands (See Fig. 1(b)). The narrow band is located at 3466 cm−1 in the IR spectrum of SLN:Fe while broad, complicated and intensive bands are present in the spectra of CLN:Fe and CLN:Fe:Hf. As the integration area of the OH- band is usually proportional to the proton concentration [15,16], the comparison of the OH- bands in Fig. 1(b) also reveals that the proton concentration is much higher in CLN:Fe:Hf than in other samples. Figure 1(c) shows the UV-Vis absorption spectra of three samples. The 500 nm-absorptions (Fe2+) of CLN:Fe:Hf and SLN:Fe are around 2 cm−1, much less than that of CLN:Fe. Figure 1(d) shows the PILS experiment setup. The e-polarized 532 nm light beam (Cnilaser) with a diameter of 2 mm impinged on the samples along the y axis. An aperture was used to block the wide-angle scattering light. Besides, a reference beam was used for preventing the laser power fluctuation. Thermal equilibration was guaranteed before each measurement. All measurements were performed at one place of samples to avoid the influence of sample unhomogeneity on the PILS measurements. The photorefractive damage was erased at 200 °C.
Fig. 1 Raman spectra (a), OH− absorption bands (b) and UV-VIS spectra (c) of CLN:Fe, CLN:Fe:Hf and SLN:Fe. (d) Experimental arrangement for PILS measurement.
3. Results and discussions
Normalized transmitted beam intensity as a function of exposure time (called as PILS curve hereafter) at three samples is plotted in Fig. 2(a). The typical PILS curve includes decreasing stage (scattering increasing) and partial recovery stage (scattering decreasing). The initial transmittance P0 decreases quickly to the minimum value during the time of τ1, and then gradually increases to a steady-state value P2 (slightly smaller than P0). The PILS curves of all three samples follow the similar evolution trends. CLN:Fe:Hf and SLN:Fe exhibit weaker scattering (higher P1 values) as compared with CLN:Fe. Some anomalous spikes appear in the recovery stage of CLN:Fe. These spikes were explained by Augustov et al. by the intermittent recombination of photoinduced space charges through the surface conductivity [17]. We measured the PILS curves of CLN:Fe at varied temperatures ranging from 20 to 200 °C, as well as the “liquid nitrogen temperature” (see Fig. 2(b)). It is found that the spikes appear only in the moderate temperature range, i.e. they are absent at −190 °C and above 120 °C. These spikes are never observed in CLN:Fe:Hf and SLN:Fe crystals, whose bulk conductivities are supposed to be higher than CLN:Fe. The noticeable result is the temperature dependence of the evolution trend of PILS curve (see Fig. 2(b)). The self-recovery process found at room temperature disappears at −190 °C or above 120 °C. Zhang et al. [11] explained the partial self-recovery process through the competition between the dc and ac components of refractive index perturbation. But the physical essence of this competition has to be clarified because no partial self-recovery process was observed at −190 °C or above 120 °C for all samples.
Fig. 2 (a) PILS curves for three samples at room temperature. (b) PILS curves for CLN:Fe at different temperatures. (c) The illumination was paused during period from 2280 to 2418s in CLN:Fe under air.
It could be found for CLN:Fe that, the anomalous spikes and self-recovery process share a similar temperature dependence. In other words, anomalous spikes and self-recovery process co-exist in the PILS curves at 20 and 100 °C while they disappear simultaneously from the PILS curves at −190 °C and above 120 °C. The correlation between the anomalous spikes and self-recovery process reveals that they may have the similar physical origin: the conductivity. Anomalous spikes are caused by the intermittent recombination of photoinduced space charges through the surface conductivity, while the self-recovery process is possibly induced by the slow recombination of photoinduced space charges through the bulk conductivity. Figure 2(c) shows the PILS curve where the illumination was paused for 138s. It can be seen that the recovery of transmittance still continues in the dark, manifesting that this effect is indeed related to the charge recombination. The charge recombination rate depends not only on the bulk conductivity but also on the local space charge field. After the illumination is turned on back, the space charge field is re-established and the change of the recovery rate can be observed. Noise gratings with different wave vector K may build up inside crystal because of the interference of the main beam and the scattering lights. The refractive index perturbation includes Δn0(t,x) and Δn1(t,x)cos(Kx + ϕ), where the term Δn0(t,x) and Δn1(t,x)cos(Kx + ϕ) are caused by dc and ac components, respectively [11]. And Δn0(t,x) + Δn1(t,x)cos(Kx + ϕ)< = Δnsat has to be satisfied. Δnsat is the saturation refractive-index modulation. At the beginning of PILS, noise gratings build up with no competition, corresponding to the decreasing stage of the transmittance. After the total refractive index perturbation reaches Δnsat, the competition between Δn0(t,x) and Δn1(t,x)cos(Kx + ϕ) starts. As compared to the dc component Δn0(t,x), the ac component Δn1(t,x)cos(Kx + ϕ) has shorter charge space (i.e. shorter period Λ) and the charge recombination happens more easily to them. As a result, the dc component Δn0(t,x) becomes dominant in the competition. Noise gratings with high wave vectors are gradually erased while the transmittance of the main beam recovers partially. As shown in Fig. 2(b), this self-recovery process is quite pronounced at the temperature of 20 and 100 °C. At −190 °C, the charge recombination is suppressed significantly due to the super low bulk conductivity of crystal. Thus, the self-recovery process is absent at the low temperature. Above 120 °C, the refractive index perturbation sum hardly reaches Δnsat because the photorefraction is strongly suppressed by the tremendous bulk conductivity. Therefore, both the competition and self-recovery process disappear at the high temperature.
The bulk conductivity responsible for the slow recombination of photoinduced space charges usually has two sources: the electron tunneling between Fe sites and the proton compensation. Yang et al. studied the effect of electronic and ionic conductivity to the space charge field in LN:Fe crystals, and found at low temperatures electron tunneling prevails while upon heating the mobility of the protons is increased significantly and migrating protons dominate the conductivity [15]. By linear fitting, the activation energies for the electronic and ionic conductivity are found to be 0.28 and 1.0 eV. The cross point of the two fitting line reveals that the temperature for the dominance transition is around 85 °C, indicating the protons start to be activated at this temperature [15]. The Fe doping concentration of 0.03 wt% in our work is much lower than that in Ref [15]. But the proton activation temperature (85°C) will not be influenced too much by the Fe doping level. The detailed effect of the doping defects (i.e. Fe2+/3+, Hf4+) on the behavior of the protons needs further investigations. It should be noted that the material parameters such as the refractive index and electro-optic coefficient may also vary with the increasing temperature. However, the enhancement of both the proton mobility and the bulk conductivity at high temperature [16] produces the most remarkable influence to the photorefractive process in LN:Fe. This can be proved by the important role of protons in the thermal fixing process of holograms [16]. Thus, the temporal dynamics of PILS in our cases may also be affected by the activation of protons.
The temperature and intensity dependences of PILS curves in SLN:Fe and CLN:Fe:Hf (see Fig. 3) show the vanishing of self-recovery process. For SLN:Fe at the room temperature, the transmittance decreases with the increasing pump intensity, because PILS usually becomes serious at higher pump intensity. But this is not valid for the case of CLN:Fe:Hf, in which the transmittance does not change monotonically with the increasing pump intensity. Zhang et al. studied the pump intensity dependence of PILS in codoped LN:Fe, and found the R value (represent the ratio of the PILS intensity to the pump intensity) changes with the increasing pump intensity in a complicated way [18]. In Fig. 3, self-recovery process vanishes gradually with the increasing temperature in CLN:Fe:Hf. However, there is a “scattering acceleration effect” in the temperature range of 60~100 °C (see Fig. 3(c)-(f)), i.e the time τ1 for the beam transmittance reaching the minimum value P1 is shortened remarkably in this range. This effect could be seen through the comparison of τΑ and τΒ, where A and B points represent the minimum transmittance for 100 and 60 °C, respectively. As the temperature range of 60~100 °C coincides with the proton activation temperature (85 °C), this “scattering acceleration effect” could be connected with the proton activation. Assuming some protons distribute by chance around Fe2+/3+, these positive-charged protons may have the tendency to localize the electrons and in turn reduce the quantum efficiency for electron excitation. Once the protons are activated, the photoexcitation of Fe2+/3+ will be no longer influenced by the migrating protons. As a result, the crystal photoconductivity increases, and the PILS process is accelerated. As a matter of fact, the contribution of “hot electrons” to the space-charge transport in LN:Fe was reported by Goulkov et al [5–7]. The “hot electrons” may be easily influenced by the surrounding impurity ions such as protons. From this view of point, it is reasonable that the activated protons cause the “scattering acceleration effect”.
Fig. 3 PILS curves at different temperatures and pump intensity for (a)-(c)SLN:Fe and (d)-(f) CLN:Fe:Hf.
Different from the results of CLN:Fe:Hf, the “scattering acceleration effect” only occurs at high intensity in SLN:Fe crystal. This is due to the large difference of proton concentrations in both crystals. A great deal of impurity defects such as HfLi and HfNb may form in the lattice of CLN:Fe:Hf [19], and these defects make protons more easily incorporated into the lattice. By contrary, fewer defects are included in SLN:Fe and lower concentration of protons are present in its lattice, which can be proved by the sharp narrow OH- peak at 3466 cm−1. In this case, the probability of the protons distributing around Fe sites is quite low. If few Fe3+/2+ are photoexcited, the influence of protons on the photoexcitation of Fe3+/2+ will be weak. Therefore, the “scattering acceleration effect” is absent at low intensity in SLN:Fe.
According to Ref. 19, the Hf concentration in CLN:Fe:Hf should be 4.5 mol%. Usually, heavy codoping may largely influence the Fe3+/2+ trap density because most of Fe ions at the Li sites are repelled into the Nb sites. However, Li et al. found that Fe ions still remain at Li sites even at high Hf doping level [14]. Therefore, the Fe3+/2+ trap density cannot be influenced too much by the Hf doping. Although the Fe3+/2+ trap density in three samples is similar, the absorptions of the samples (see Fig. 1(c)) reveal the [Fe3+]/[Fe2+] ratio varies with the Hf-doping and the Li incorporation. In our case, the [Fe3+]/[Fe2+] ratio is much lower in CLN:Fe than in other samples. Goulkov et al. studied the photoelectric response in LN:Fe versus the [Fe3+]/[Fe2+] ratio in Ref. 7, and showed that the photovoltaic field Epv is not only proportional to [Fe3+] but also inversely proportional to the term (eμϕs)/(γhνv), where μ, ϕ, s, γ, hν, and v are the electron mobility, the quantum efficiency, the photoionization cross section, the recombination cross section, the photon energy and the electron velocity, respectively. The Li incorporation or the Hf-doping can enhance the LN photoconductivity through removing the intrinsic defects such as NbLi ions [12–14], and this enhancement is related to (eμϕs)/(γhνv) rather than the [Fe3+]/[Fe2+] ratio. As a result, Epv is suppressed by the Hf-doping or the Li incorporation in CLN:Fe:Hf or SLN:Fe. In Ref. 5-7, Goulkov et al reported the competition between the photovoltaic field Epv and the diffusion field Ed. Since Epv remains constant while Ed goes up quickly with the increasing wave vector K, the noise gratings with high wave vectors are more connected with Ed than with Epv. In this view, the competition between the noise gratings with different wave vectors could be roughly considered as the competition between Epv and Ed. Since Epv is quite lower in SLN:Fe than in CLN:Fe, the diffusion field Ed should govern all the photorefractive processes (including PILS) in SLN:Fe. Higher coupling gain but lower diffraction efficiency was found in SLN:Fe than in CLN:Fe [12], which reveals the dominant role of Ed in SLN:Fe. Consequently, the PILS is expected to be more asymmetric in SLN:Fe. Since Ed goes up with the increasing wave vector K, the scattering intensity should tend to distribute toward the wide-angle range in SLN:Fe. For clarifying the competition between the diffusion and photovoltaic effects, the angle distribution of scattering intensity, just like the precise works of Goulkov et al., has to be preformed.
The self-recovery process was explained by steady-state temperature gradient generated by focused laser in the work of Kostritskii et al [10]. Basing the systematic SCF (SCF = 1-Δnsat/Δnmax, describing the partial compensation degree of the optical damage) and PILS kinetics comparison, they interpreted the relevant mechanisms as pyroelectricity and the electron-hole competition in photoinduced charge transport. In Ref. 10, the used intensities were in the range of 3.9~363.6 W/cm2, but the intensities (unfocused laser) in our case are in the range of 0.076~0.228 W/cm2. Such lower intensities cannot induce sufficient steady-state temperature gradient in the sample. CLN:Fe was also studied in Ref. 10, but no self-recovery process was reported. This is possibly due to the focusing of the laser beam. As explained in Ref. 20, the PILS induced by focused laser is restricted only to the gratings with low wave vectors, and therefore it is more connected with photovoltaic effect than with diffusion effect. Thus, the competition between the noise gratings with different wave vectors could be totally different from our case, and the self-recovery process is absent from their cases.
4. Conclusion
We studied the self-recovery processes in CLN:Fe, CLN:Fe:Hf and SLN:Fe crystals. The self-recovery process was explained by the slow recombination of photoinduced space charges of the noise gratings with high wave vectors. The “scattering acceleration effect” was connected with the activated protons located around Fe ions and their impact on the photoconductivity. The low concentration of protons in near-stoichiometric SLN:Fe crystal was suggested to account for the relationship between the pump-intensity and “scattering acceleration effect”.
Acknowledgments
We thank the referees for their valuable comments. This work is supported by the National Natural Science Foundation of China, No.61108060, YQ2013029, 212016, F2013202153 and CG2013003002.
References and links
1. I. N. Kiseleva, V. V. Obukhovskii, and S. G. Odoulov, “Parametric scattering of the holographic type in class 3m crystals,” Sov. Phys. Solid State 28, 1673 (1986).
2. M. Segev, D. Engin, A. Yariv, and G. C. Valley, “Temporal evolution of fanning in photorefractive materials,” Opt. Lett. 18(12), 956–958 (1993). [CrossRef] [PubMed]
3. H. Rajbenbach, A. Delboulbé, and J. P. Huignard, “Noise suppression in photorefractive image amplifiers,” Opt. Lett. 14(22), 1275–1277 (1989). [CrossRef] [PubMed]
4. M. A. Ellabban, G. Mandula, M. Fally, R. A. Rupp, and L. Kovacs, “Holographic scattering as a technique to determine the activation energy for thermal fixing in photorefractive materials,” Appl. Phys. Lett. 78(6), 844 (2001). [CrossRef]
5. M. Goulkov, S. Odoulov, Th. Woike, J. Imbrock, M. Imlau, E. Krätzig, C. Bäumer, and H. Hesse, “Holographic light scattering in photorefractive crystals with local response,” Phys. Rev. B 65(19), 195111 (2002). [CrossRef]
6. M. Goulkov, M. Imlau, and Th. Woike, “Photorefractive parameters of lithium niobate crystals from photoinduced light scattering,” Phys. Rev. B 77(23), 235110 (2008). [CrossRef]
7. M. Goulkov and Th. Woike, “Photoelectric response in LiNbO3:Fe versus Fe2+/Fe3+ ratio studied by PILS method,” J. Opt. Soc. Am. B 31(5), 1071–1077 (2014). [CrossRef]
8. S. M. Kostritskii and O. G. Sevostyanov, “Influence of intrinsic defects on light-induced changes in the refractive index of lithium niobate crystals,” Appl. Phys. B 65(4-5), 527–533 (1997). [CrossRef]
9. S. M. Kostritskii, O. G. Sevostyanov, M. Aillerie, and P. Bourson, “Suppression of photorefractive damage with aid of steady-state temperature gradient in nominally pure LiNbO3 crystals,” J. Appl. Phys. 104(11), 114104 (2008). [CrossRef]
10. S. M. Kostritskii, M. Aillerie, and O. G. Sevostyanov, “Self-compensation of optical damage in reduced nominally pure LiNbO3 crystals,” J. Appl. Phys. 107(12), 123526 (2010). [CrossRef]
11. X. Zhang, J. Xu, S. Liu, H. Huang, J. Wolfsberger, X. Chen, and G. Zhang, “Temporal evolution of beam fanning in LiNbO3:Fe, In crystals,” Appl. Opt. 40(5), 683–686 (2001). [CrossRef] [PubMed]
12. Y. Furukawa, K. Kitamura, Y. Ji, G. Montemezzani, M. Zgonik, C. Medrano, and P. Günter, “Photorefractive properties of iron-doped stoichiometric lithium niobate,” Opt. Lett. 22(8), 501–503 (1997). [CrossRef] [PubMed]
13. H. Liu, X. Xie, Y. Kong, W. Yan, X. Li, L. Shi, J. Xu, and G. Zhang, “Photorefractive properties of near-stoichiometric lithium niobate crystalsdoped with iron,” Opt. Mater. 28(3), 212–215 (2006). [CrossRef]
14. 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]
15. Y. Yang, I. Nee, K. Buse, and D. Psaltis, “Ionic and Electronic Dark Decay of Holograms in LiNbO3:Fe Crystals,” Appl. Phys. Lett. 78(26), 4076 (2001). [CrossRef]
16. K. Buse, S. Breer, K. Peithmann, M. Gao, and E. Krätzig, “Origin of thermal fixing in photorefractive lithium niobate crystals,” Phys. Rev. B 56(3), 1225–1235 (1997). [CrossRef]
17. P. A. Augustov and K. K. Shvarts, “Surface Recombination and Photorefraction in LiNbO3:Fe Crystals,” Appl. Phys. (Berl.) 18(4), 399–401 (1979). [CrossRef]
18. G. Zhang, G. Zhang, S. Liu, J. Xu, Q. Sun, and X. Zhang, “The threshold effect of incident light intensity for the photorefractive light-induced scattering in LN:Fe,M (M=Mg, Zn, In) crystals,” J. Appl. Phys. 83, 4392 (1998). [CrossRef]
19. S. Li, S. Liu, Y. Kong, D. Deng, G. Gao, Y. Li, H. Gao, L. Zhang, Z. Hang, S. Chen, and J. Xu, “The optical damage resistance and absorption spectra of LN:Hf crystals,” J. Phys. Condens. Matter 18(13), 3527–3534 (2006). [CrossRef]
20. H. Qiao, J. Xu, G. Zhang, X. Zhang, Q. Sun, and G. Zhang, “Ultraviolet photorefractivity features in doped lithium niobate crystals,” Phys. Rev. B 70(9), 094101 (2004). [CrossRef]
