1. Introduction

Lithium niobate (LN) is an attractive material widely used in integrated optics community [1]. Proton exchange (PE) [2–10] was known as a basic technology for fabricating index structure on LN substrate. It is based on a reaction of Li-H ion exchange of LN in a suitable acidic source, and the protons incorporation can confine the light propagation in the crystal through modulating the material index [2–4], which makes PE technology applicable to photonics. In the past few decades, Korkishko and Fedorov et al. [4–6] carefully studied the PE process of LN and established a correlation between the crystal structure and its refractive indices. At the mean time, Cabrera and his colleagues made a throughout investigation on the fabrication and characterization of the PE waveguide [3, 7–9], for example, Carnicero and Carrascosa et al. [8, 9] found the intensity thresholds for optical damage in α-phase PE LN waveguides could be increased from the substrate value to a factor 500 greater by increasing the PE duration. Recently, the EO properties of the PE waveguides were reported by Kostritskii et al. to depend on the phase composition changed intentionally by varying fabrication conditions [10]. All these works emphasized on how the light propagation is appropriately controlled by the experimental parameters of the PE and subsequent treatments. However, a question may be raised here: could the PE process be confined or influenced by the light?

Nowadays, most of photonics structures are fabricated through photolithography technique. The same situation applies to the PE-based cases and selective PE can only be achieved through this complicated and time-consuming route. However, some recent researches show that photo-assisted processes may contribute to the easy preparation of the photonics structure. Selective domain inversion and spatially resolved superhydrophilicity were achieved by different groups on the LN surface under light assistance [11–13]. One may think whether the selective PE process could be realized on LN under the light assistance.

Besides the material index modification, the heavy incorporation of hydrogen ions into the lattice might also induce the chemical etching (CE) of LN. Some researchers reported that proton first reacts with Li ion on the crystal surface and this plays a crucial role in the subsequent etching process [14]. It is well-known that in acid the –z face appears to be etched easily whereas the + z face remains hardly touched. However, does this rule still hold true when the laser irradiation is introduced to the PE and the subsequent CE processes?

In this paper, the photo-assisted proton exchange (PAPE) and the subsequent chemical etching (PACE) will be performed in Fe-doped LN crystals. It will be shown that PE and CE are selectively controlled by light irradiation. Additionally, some special etching results that are connected with photogalvanic effect (PVE) will be reported.

2. Experimental procedures

The samples used in our experiments were 0.7 mm-thick, z-cut, Fe-doped congruent lithium niobate (CLN) crystals and the doping concentration is 0.03 mol%. In order to modify the ratio of Fe_Li²⁺/ Fe_Li³⁺ (this ratio can be characterized indirectly through the crystal absorption) the samples were annealed in vacuum (10⁻¹Pa) at different temperatures (500-1000 °C). Figure 1(a) shows the absorption spectrum of Fe-doped LN samples studied in our work. As we know, the electron traps Fe_Li²⁺ in LN may induce strong wide-band absorption around 2.5eV (500nm). Under the illumination, the electrons could be easily photo-excited from Fe^2+/3+ traps, forming the current toward the –z surface (i.e. photogalvanic effect).

 

Fig. 1 (a) UV-VIS spectrum of Fe-doped LN samples used in Case 1, 2, 3 and 4. (b) Outline of the experimental setup for PAPE and PACE.

Download Full Size | PDF

Our experiment involves both the proton exchange of the sample in pyrophosphoric acid (PSA) and the exposure of the sample to the 455nm-laser (CW) radiation at the same time. As shown in Fig. 1(b), the laser beam with a power of about 1.2W was focused on the sample surface using a lens with 300mm focal length and the focused spot has a size of about 500μm. Then the sample was mounted in a quartz holder and placed into a vessel containing PSA. After the PSA temperature was raised rapidly to 60 °C, the sample was exposed to the focused beam instantly. Our preliminary experiment shows PSA at the low temperature of 60 °C cannot induce obvious PE effect alone, and the hydrogen incorporation is due to the combined effect of PE and the laser irradiation. By subtracting the reflectance of the vessel window and the sample surface (total ~35%) and taking into account the absorption coefficient (~40 cm⁻¹) and the thinness (0.7 mm) of the sample, we could estimate the intensity at the focus spot to be 400W/cm² and the energy absorbed by the material to be 0.73J per second. If only the heat conduction inside the crystal (LN has a low thermal conductivity) is considered, such an absorbed energy may lead to the temperature at the focus ramping to several hundreds of centigrade in few minutes. However, the direct heat exchange between the crystal and acid at the focus and its surrounding region significantly lower the local temperature. The majority of heat dissipates through acid convention into the environment. Though it is difficult to measure the temperature at the focus experimentally, we could estimate its range by irradiating the sample under the same condition but in the liquid with different boiling point, for example, water (100°C) or oil (~200°C). What we observed is that the tiny boiling bulbs are produced continuously near the focus when the sample was irradiated in the water but nothing happens in the oil. It indicates that the average temperature of the focus region is likely in the range of 100~200 °C. In fact, the LN surface undergoing a long irradiation in oil was carefully checked by optical microscope and no damage such as micro-cracks was found.

After the treatment of the desired time, the sample was cleaned for characterization. Stand-alone FT-IR microscope (Bruker LUMOS) was employed to analyze the spatial distribution of proton concentration. It is well known that a narrow band centered at around 3500 cm⁻¹ in the IR spectra (Figs. 2(a), 2(c) and 2(e)) is associated with the O-H stretching vibration [3]. For each sample, 100 points (10 × 10 lattice) were selected around the surface area undergoing the PAPE process, and the FT-IR microscope collected point by point the IR spectrum with a resolution of 1 cm⁻¹ and a scan time of 32 s/point. In order to obtain the spatial distribution of the relative proton concentration, the integration of the O-H vibration band (i.e. band area) at each point was calculated and the band area proportional to the proton concentration was plotted as function of spatial axis (x and y-axis) in 2D and 3D (Figs. 2(b), 2(d) and 2(f)). Optical microscope (Olympus STM6) was used for the morphology observation (See Fig. 3).

 

Fig. 2 Micro-IR spectroscopy results reflecting the spatial distribution of the relative proton concentration after PAPE and PACE. a) and b) corresponds to Case 1, c) and d) to Case 3, and e) and f) to Case 4. In a), c) and e), the 100 points (10 × 10 lattice) denote the sample region where the FT-IR spectrum was collected, and the O-H vibration bands shown in the insets are collected respectively from the region points labeled 1 to10. In b), d) and f), the integration of the O-H vibration band is plotted as function of spatial axis (x and y-axis) in 2D and 3D.

Download Full Size | PDF

 

Fig. 3 Topographic images of photo-assisted chemical etching on a) the –z surface in Case 1, b) the + z surface in Case 2, c) the –z surface in Case 3 and d) the + z surface in Case 4. The insets are the back surfaces of these samples in Case 3 and 4. e) The cross-section of the deep etching hole on the + z surface in Case 4.

Download Full Size | PDF

3. Results and discussion

The specific conditions of the experiments were described in Table 1. The effect of PAPE process depends on these experimental conditions, especially on the absorption and the orientation of the samples ( + z or -z). The samples in Case 1 and 2 have much less absorption and experience treatments of shorter duration (8h). As shown in Fig. 2(a) (more clear in Fig. 3(a)), after PAPE the –z surface (Case 1) remains untouched except some scratches. However, the spatial distribution of the relative proton concentration (Fig. 2(b)) shows that the proton incorporation happens intensively at the laser-irradiated region while quite few protons are detected outside this region, indicating that the PE process can be controlled selectively by the laser spot and that the laser irradiation has an effect of accelerating proton exchange. After the same PAPE treatment, the + z surface (Case 2) shows a similar proton distribution profile (not shown in Fig. 2) but an etching-like sign is obvious on the + z surface (Fig. 3(b)).

Table 1. Treatment conditions and parameters

View Table

As already mentioned in the introduction, the heavy incorporation of protons may induce the LN surface etching. The conventional etching (i.e. etching in the dark) is quite anisotropic: it happens easily on the –z surface of LN (hardly on the + z surface) even in the strong acid such as HF. In fact, our preliminary experiment shows that the heavy conventional PE treatment (i.e. PE in the dark) at a high temperature above 230 °C indeed results in a slight etching on the –z surface of LN but nothing on the + z surface. In the subsequent experiments, we will study whether the PAPE induces such kind of surface etching.

Since we suggest above that the laser irradiation might enhance the extent of the proton exchange, the Fe-doped LN with larger absorption is selected in Case 3. Additionally, the proton exchange duration is prolonged to 12h. It can be seen from Fig. 2(c) and Fig. 3(c) that a round area, which coincides to the laser spot, is obviously etched after PAPE. Moreover, the spatial distribution of the relative proton concentration plotted in Fig. 2(d) shows a Gaussian-like peak. These results, on one hand, confirm our above suggestion that the laser irradiation could enhance the proton exchange and make this process controllable spatially, on the other hand, reveal that the heavy PAPE may induce surface etching and transform to PACE.

The + z surface (Case 4) is attempted under the same experimental conditions, and the PACE effect is also found on the + z surface (see Fig. 2(e) and Fig. 3(d)). The noticeable result in Fig. 2(e) is that a deep etching hole appears in the center of irradiated area. Figure 3(e) shows the detail of the etching hole forming in Case 4: it usually has a wedge-like shape and a depth of hundreds of microns. In Figs. 3(c) and 3(d), we compare the etching topography of the irradiated surface of Case 3 and 4. Note that, like the Case 1 and 2, the same experimental conditions were also adopted here except the different irradiated surfaces ( + z or -z). For the -z surface, the etching extent varies gradually from the spot center to the surroundings, while on the + z surface the etching extent suddenly increase wildly at the center forming a deep hole there. This result reinforces the comparison result between Case 1 and 2, indicating that the + z surface tends to be more etched under irradiation than the –z surface, which is quite different from the etching resistance demonstrated by the + z surface in the conventional LN etching experiments. We also examine the back surface (i.e. the opposite one of the irradiated surface) of the samples in Case 3 and 4 (see the insets of Figs. 3(c) and 3(d)). Even though the irradiated surface ( + z surface) in Case 4 undergoes a heavier etching than that (-z surface) in Case 3, the back surface (-z surface) in Case 4 demonstrates some unexpected etching resistance as compared with that ( + z surface) in Case 3. This phenomenon is consistent with the above comparison results, revealing that the proton exchange (or proton incorporation) happens more easily on the + z surface when the laser irradiation is introduced to the conventional PE or CE process.

Regarding the mechanism of the PAPE and PACE, the thermal effect induced by the material absorption to the blue light should be considered first. In the illuminated region, the laser irradiation behaves as a volume heating source and induces the local temperature raise, which enhances the proton exchange and the subsequent chemical etching in the illuminated region. However, the thermal effect cannot explain entirely the results, especially the unexpected etching difference between + z and –z surfaces. This etching anisotropy under the laser irradiation contradicts the previous conclusion drawn from the conventional etching experiment, and it should be connected with the photo-related anisotropy in the Fe-doped LN—–photogalvanic effect (PVE). The mechanism related with PVE is given in Fig. 4. Under the irradiation, the photo-excited electrons in Fe-doped LN transport along one fixed direction, inducing a current toward the –z surface (i.e. PVE). This current leads to the accumulation of the electrons in the + z surface and to the congregation of the positive charged traps in the –z surface. The excessive negative charges in the + z surface may attract the protons in the acid for charge compensation, resulting in the acceleration of the proton exchange and the subsequent chemical etching. Accordingly, the excessive positive charges in the –z surface will repel the protons and induce the etching resistance of the –z surface. The most obvious difference between the two comparison groups (Case 1, 2 and Case 3, 4) is on the sample absorption (18 and 42 cm⁻¹). The results show that the etching anisotropy of Case 3 and 4 is much stronger than that of Case 1 and 2, indicating the sample absorption could enhance this etching anisotropy. Considering the photogalvanic current usually increases with the sample absorption, the PVE contribution to the PACE can be further confirmed. Note that irradiating the LN surface intensively in air often causes the over high temperature at the focus and the degradation of the local crystal. This behavior is also dependent on the sample absorption. However, it cannot explain the special etching anisotropy. Moreover, in our case irradiating LN was carried out in the liquid and the strong convention may lower significantly the local temperature. The etching enhancement due to the crystal degradation is not likely to happen.

 

Fig. 4 The PAPE and PACE mechanism related with the photogalvanic effect.

Download Full Size | PDF

4. Conclusion

PAPE and PACE experiments were attempted on Fe-doped LN crystals. Selective PE and CE can be achieved by 455nm-laser irradiation on the crystal surface in PSA. The Micro-IR spectroscopy analysis show that the hydrogen incorporation is confined spatially by the laser irradiation. Moreover, under the laser irradiation, + z surface is found to be more easily etched than –z surface, which is explained by the photogalvanic effect of the Fe-doped LN.