Direct UV laser writing on chromium coated lithium niobate (LiNbO3) crystals is found to produce spontaneous domain inversion associated with the exposed UV laser tracks. Experimental evidence suggests that this effect is attributed to local out-diffusion of oxygen, reducing the LiNbO3 crystal surface due to the presence of chromium. The thin chromium film becomes hot and reactive after absorbing the UV laser radiation thus acting as an oxygen getter. This very efficient process enables the inversion of domains at lower intensities as compared to other direct laser based poling methods practically eliminating the deleterious surface damage induced by the direct absorption of the UV laser radiation by the crystal. Furthermore, the versatility of this domain fabrication method, is demonstrated by the production of inverted domain structures on Z-, Y- and 128°YX-cut substrates.
© 2014 Optical Society of America
Domain engineering in lithium niobate (LiNbO3) crystals, especially for the fabrication of periodic domains, is an attractive field of research with its most prominent applications in nonlinear optical frequency conversion [1–3]. Periodic domain inversion is usually achieved using ad hoc electric field poling, by which a spatially modulated electric field is applied along the Z-axis of a crystal [4,5], exceeding the coercive field . The resultant domain pattern depends on the pattern of the insulating dielectric coating on the crystal surface, which is photolithographically defined prior to the poling. While electric field poling has been proven to be an effective technique for a wide range of applications [1–3], it becomes impractical for non-polar crystal cuts such as X- and Y-cuts or even on a crystal with an inclined polarization axis such as 128° YX-cut LiNbO3 [7,8]. However, periodically poling on X-cut LiNbO3 would enable the integration of frequency converters with electro-optical modulators, which benefit from X-cut for temporal stability . Also, on 128° YX-cut LiNbO3, periodic poling would enable the effective generation of surface acoustic waves (SAW) making use of the relatively large electromechanical coupling coefficient .
Other domain engineering techniques, such as titanium (Ti) in-diffusion, have been reported to achieve domain inversion on cuts with an inclined polarization axis [11,12]. However, this high temperature process (~1040°C) needs a carefully controlled environment and cannot be employed as a post-processing step. Recently, the fabrication of domains by local UV illumination has been demonstrated to be applicable also to non-polar LiNbO3 crystal surfaces [13–16]. Although encouraging, this technique has two drawbacks: a severe limitation of the domain depth to only a few microns; and obstructive, thermally induced surface damage. Since the depth of the achieved domains is related to the heat deposited at the surface by the UV laser light, the choice of UV intensity to use for domain writing represents a trade-off between the achievable domain depth and the degree of surface damage. In order to circumvent this limiting trade-off, coating the crystals with a thin metallic layer before UV irradiation seems to be a natural next step towards the fabrication of domain patterns with reduced surface damage.
In this paper, we report on domain engineering of LiNbO3 by coating the crystal with a thin metal layer (chromium) prior to UV laser-beam irradiation. The inverted domains on Y-, Z- and 128° YX-cut LiNbO3 have been analyzed for various intensities of UV laser light, thicknesses of the Cr-layer, and different processing atmospheres, namely air and dry nitrogen. An analysis of the achieved domain depth and quality is presented and the possible physical mechanisms leading to the observed domain inversion behavior are discussed.
2. Experimental methods
The studies were performed with Y-, Z- and 128° YX-cut congruent LiNbO3 crystals. The crystals were coated with 20, 40 or 60 nm of chromium (Cr) thin films that were deposited by e-beam evaporation. Control experiments were carried out on 40 nm platinum coated and virgin LiNbO3 crystals to compare the performance of domain writing of Cr coated crystals. UV irradiation for domain formation was performed using a continuous wave, frequency-doubled argon ion laser (λ = 244 nm). The laser beam was tightly focused with a fused silica lens (focal length f = 40 mm) to a focal beam diameter of ~7 μm. The irradiation intensity was varied between 1.75 × 105 and 2.92 × 105 W/cm2. In order to enable control of the atmosphere of the domain writing (ambient air or dry nitrogen), the crystal was positioned within a closed chamber containing a fused silica window through which light was focused onto the crystal. The chamber was mounted on a three-axis computer-controlled translation stage. Laser tracks could be written onto the crystals in any predefined direction (velocity: 0.1 mm/s).
The optical properties of the Cr coating and the LiNbO3 crystal with regard to the UV illumination can be summarized as follows: the reflectivity of the Cr thin film is ~40% , whereas uncoated LiNbO3 reflects only 28% of the incident light . The absorption length is ~10 nm for Cr  and ~30 nm for LiNbO3 .
The Cr coated samples were inspected for any visual surface modifications before and after stripping off the Cr layer with a Cr etchant using an optical microscope. In order to analyze the generated domains, the different samples were diced and polished as follows: i) The Y-cut crystals were diced orthogonally to the UV tracks and polished to optical grade; ii) The Z- and the 128° YX-cut crystals were wedge polished at an angle of ~6°, thereby stretching the depth profile by a factor of ~10. These two different approaches were used in order to exploit the high etching rate differential on the Z face of the crystal . For the visualization of the domains generated by the UV irradiation, all crystals were HF etched followed by SEM imaging. The stability of the UV written domains was tested by a three hour heat treatment on a hot plate at various temperatures.
3. Experimental results and discussion
3.1 Optical inspection of the Cr-coated Z-cut LiNbO3 crystals irradiated with UV-light
As a first experiment, the UV irradiated tracks of the Cr-coated crystals were simply inspected using optical microscopy to assess any visible change in the appearance of the Cr layer itself. We prepared a specific sample for this purpose by irradiating an area of 2 × 2 mm2 by scanning 200 lines of 2 mm length with a line spacing of 10 µm on a 40 nm Cr coated Z-cut LiNbO3 crystal (see inset in Fig. 1(b)). The UV writing was performed with intensities of 1.75 × 105 and 2.92 × 105 W/cm2 in ambient air and in dry nitrogen atmosphere.
The area of the UV-irradiated Cr-layer could be easily distinguished from the surrounding area since for all intensities used, it showed a rougher topography and the Cr had turned from metallic to green. This color change is more pronounced under ambient conditions (air). After removing the Cr layer, we observed that there was a change of color of the UV irradiated areas of the crystal. As can be seen in Fig. 1, the laser-irradiated area had darkened with respect to the surrounding area for all four cases. However, the strongest effect by far can be seen for an intensity of 2.92 × 105 W/cm2 in conjunction with nitrogen atmosphere: the LiNbO3 crystal became very dark.
The green color of the Cr-layer after UV-irradiation indicates oxidization of the Cr to form Cr2O3 . The formation of Cr2O3 can be described as follows: Upon heating the Cr layer with the irradiating laser beam, Cr becomes reactive, thus reacting with the oxygen from the ambient air atmosphere .
The color change of crystal itself when using an intensity of 2.92 × 105 W/cm2 in nitrogen atmosphere, Fig. 1(d), was surprising, since it is drastically different to the color change of the crystal when using the same intensity in an ambient atmosphere, Fig. 1(b). The dark color of the crystal in Fig. 1(d) indicates that oxygen diffused out of the crystal, leading to an oxygen deficiency at the surface [22,23]. The out-diffused oxygen most likely reacted with the hot Cr layer, which acts as an oxygen sink, to form Cr2O3.
The whole process, which is proposed to explain the results in Fig. 1 is illustrated in Fig. 2, where the upper row sketches the Cr oxidation process in ambient atmosphere whereas the lower row shows the process in dry nitrogen atmosphere. In Fig. 2(b) the irradiated Cr reacts with the oxygen from the atmosphere and the oxygen that becomes mobile from the LiNbO3 crystal to form Cr2O3 on both of the Cr interfaces. The oxygen in the air is expected to be more mobile than the oxygen in the LiNbO3 crystal, where oxygen out-diffusion starts to take place at temperatures of 500-600°C [23,24]. Therefore, the Cr2O3 formation on the air-Cr interface should be faster than on the Cr-LiNbO3 interface. For higher UV light intensities this process is even accelerated, which would result in a completely oxidized Cr layer as it is illustrated in Fig. 2(c). In this the Cr layer is shown to be fully depleted and cannot incorporate any more oxygen.
The dry nitrogen atmosphere in Fig. 2(d) does not contain any oxygen, hence, irradiating the Cr coating in this atmosphere would only result in Cr2O3 formation on the Cr-LiNbO3 interface, as it is illustrated in Fig. 2(e). For higher UV light intensities the complete oxidization of the Cr layer can therefore be delayed because the only source of oxygen is lithium niobate. More oxygen would diffuse out of the LiNbO3 crystal when higher intensities are used in a nitrogen atmosphere, leading to a dark crystal color as shown in Fig. 1(d). This process will be limited either by full oxidization of the Cr layer or by the duration of the UV irradiation. This suggests that it is possible to control the chemical reduction of the LiNbO3 crystal by selecting an appropriate UV laser intensity and irradiating the Cr layer in a nitrogen atmosphere.
3.2 Comparison of the domain formation in Cr coated and uncoated Z-cut LiNbO3 crystals
To investigate the influence of the Cr coating on the domain inversion process, we compared the domain formation on Cr coated and unreactive platinum coated and uncoated LiNbO3 crystals. We therefore irradiated a 40 nm Cr coated, platinum coated and an uncoated LiNbO3 crystal with UV laser intensities of 2.14 × 105 and 2.73 × 105 W/cm2, respectively, in ambient air. In each case, domain writing was performed on the + and -Z face of the crystal using a line spacing of 30 µm between the individual tracks. Domain inversion was only observed on the -Z face of the Cr coated and uncoated crystals. The Pt coated crystal did not show any evidence of domain inversion. The stability of the UV written domains formed on the Cr coated samples was tested via heat treatment at a temperature of 200°C on a hot plate for three hours and no change in the inverted domains was observed.
The SEM images of the wedge-polished and HF etched crystals are presented in Fig. 3, where a series of observations can be made: i) domain inversion can be seen in both cases; ii) the inverted domain of the Cr coated sample in Fig. 3(a) is roughly two times wider and deeper as compared to the uncoated sample of Fig. 3(b); iii) there is reduced surface damage for the Cr coated sample; iv) the shape of the domains differs with domains of the Cr coated sample of Fig. 3(a) exhibiting a winged ‘Gaussian’ shape, whereas the domains of the uncoated sample of Fig. 3(b) being rather conical; v) the domain of the uncoated crystal of Fig. 3(b) shows narrow, crossing domain-inverted features with widths of around 100 nm to both sides and below the domain's center, which are absent from the Cr coated crystal of Fig. 3 (a).
An important observation is that domain inversion for direct writing on the uncoated sample was only possible for UV-laser intensities > 2.73 × 105 W/cm2 whereas on the Cr coated sample, domain formation was possible using 2.14 × 105 W/cm2 or even lower laser light intensities as shown in Fig. 3(a).
These observations indicate that two different domain inversion mechanisms are responsible for the domain generation in a Cr coated and uncoated LiNbO3 crystals. For the uncoated crystal, both the shape of the domains and the 100 nm crossing domain-inverted features correspond to the earlier observations by Muir et al. . In this case, the thermoelectric field, induced by the temperature gradient has been proposed as an explanation of the UV domain direct writing [13,16]. As for Cr coated crystals, domain formation seems easier than for the uncoated crystals. Domains can be generated at lower UV laser intensities despite the fact that the Cr coating should reflect more UV laser light than LiNbO3.
It is know that diffusion processes such as titanium (Ti) in-diffusion or lithium out-diffusion can cause domain inversion [11,12,25–27]. These processes are usually performed in a furnace at elevated temperatures (950-1100°C). It has been reported that the concentration gradient of impurities and defects, which arises during the in-diffusion of metals (Ti) or out-diffusion of lithium, induces a built-in electric field that inverts the polarization of the crystal when exceeding the coercive field. The domain formation usually takes place at the + Z face of the crystal. Kugel et al.  suggested the domain formation when annealing an uncoated LiNbO3 crystal is due to the defect gradient when oxygen diffuses out. However, it has been shown that the domain inversion depth resulting from out-diffusion in air is deeper than that in an argon atmosphere . This indicates that the domain inversion is caused by another diffusion process. Therefore, Huang et al.  attributed the domain formation due to the out-diffusion of lithium.
In Section 3.1 we proposed that oxygen diffuses out of the surface of the LiNbO3 crystal when the Cr coating is irradiated by UV laser light, indicated by the darker color of the UV irradiated area. However, the formation of oxygen vacancies in LiNbO3 is an energetically unfavorable process , suggesting that the octahedrons in the LiNbO3 structure remain stable against oxygen vacancies. The oxygen deficiency can thus be compensated by the diffusion of cations from the surface into the crystal . This would mean that when oxygen diffuses out of the surface, a counter movement of lithium ions occurs, which diffuse into the crystal. The temperature gradient can even enhance the lithium ion movement into the crystal . Since lithium out-diffusion causes domain inversion on the + Z face, one would expect a reverse lithium out-diffusion to cause domain inversion on the -Z face. This could explain the domain inversion on the -Z face as it was found when irradiating a Cr coated LiNbO3 crystal with a focused UV laser beam.
The competing heat-driven diffusion of Cr into the crystal's surface is less probable, since the temperature for this process to be efficient is about 400°C higher (at temperatures of ~950-1050°C ) than for the out-diffusion of oxygen, which takes place at ~500-600°C [23,24] and the in-diffusion of lithium, which becomes mobile at ~400°C. Additionally, Cr reacts with oxygen at temperatures similar to those at which oxygen becomes mobile in LiNbO3 [13,14,21]. The lower temperatures at which the oxygen and lithium ions are mobile can explain why the domain inversion process in Cr coated crystals occurs at lower intensities as compared to the uncoated crystals, where crystal temperatures of ~1100°C are necessary for domain formation and the domain inversion mechanism was explained by an UV induced thermoelectric field [13,16]. We estimated the necessary temperature for domain inversion on Cr coated crystals upon UV irradiation at ~600-700°C, keeping the reflectivity of Cr and the UV intensity in mind. Verification of this estimate should be attempted, but this is deemed beyond the scope of the current work and will be attempted in a future investigation. The lower temperatures, which are necessary for domain writing in Cr coated crystals lead to the observed reduction or elimination of surface damage caused by UV irradiation, compared to uncoated crystals.
The UV domain writing process on Cr coated crystals induces also a temperature gradient in the crystal. This suggests that two sources of an induced electric field are present in the crystal when irradiating the Cr layer with UV light. Firstly, the thermoelectric field that originates from the temperature gradient and secondly the defect gradient induced electric field that originates from the out-diffusion of oxygen respectively the in-diffusion of lithium. One should also keep in mind that the reduced lithium niobate has a higher electrical conductivity , which makes a complete description of this domain inversion process even more complex. However, the findings strongly suggest that the domain formation on Cr coated crystals is due to a diffusion process, which is supported by the fact that no domain inversion was observed, when a non-reactive metal layer (Pt) was used. This can be further clarified by keeping the temperatures in mind, which are necessary for different domain inversion mechanisms. In the case for the domain inversion due to the thermoelectric effect, a temperature close to the Curie temperature of the crystal (~1100°C) and a high temperature gradient (~200°C/μm) are necessary . Domain formation due to the pyroelectric effect only needs a crystal temperature of up to 200°C . The temperatures for diffusion processes (oxygen and lithium) on the other hand are closer to the estimated crystal temperature upon UV irradiation, making the diffusion processes favorable for causing domain inversion on Cr coated crystals. However, further investigations are required in order to be conclusive.
3.3 Domain formation in Cr-coated Z-cut LiNbO3 crystals using different coating thicknesses and laser intensities under ambient conditions (air)
So far we know that UV laser irradiation of Z-cut LiNbO3 crystals, which have been coated with a thin Cr film, can cause domain inversion. In this section the influence of the Cr coating thickness and the UV laser intensity on the domain formation is investigated. UV domain writing was performed in an ambient air atmosphere on several Z-cut LiNbO3 crystals, coated with a Cr thickness of 20, 40 and 60 nm. The UV laser was scanned in linear tracks focused to achieve intensities between 2.14 × 105 and 2.44 × 105 W/cm2.Fig. 4 presents the SEM images of the domain shapes after wedge polishing and HF etching, written on different thicknesses of the Cr coating and different UV laser intensities. From Figs. 4(i)-4(l), it is clear that using a 60 nm-thick Cr layer gives entire domains under the UV irradiated track, irrespective of the intensity. Comparing the different intensities, it is evident that the domain width and depth increases with higher UV light intensities. It can also be observed that a bright stripe emerges at the surface of the domain center, when the intensity is increased. In the case of a thinner, 40 nm-thick Cr layer, a similar domain shape to the 60 nm Cr coating can only be observed at the lowest intensity of 2.14 × 105 W/cm2, Fig. 4(e). At an intensity of 2.24 × 105 W/cm2, shown in Fig. 4(f), a ‘hollow’ domain can be observed, where hollow describes a domain with an un-inverted center at the surface. At even higher intensities, as shown in Figs. 4(g) and 4(h), the domain inversion is only observed at the edges of the irradiated tracks, with the region directly under the track un-inverted. In the center of the inverted domain in Figs. 4(g) and 4(h) a shallow domain inversion part can be observed. The central shallow domain of Fig. 4(h) clearly exhibits domain-inverted features with widths of around 100 nm similar to those evident in the domains written on the uncoated samples in Fig. 3(b). For the 20 nm Cr coating, domains are only evident on either side of the irradiated tracks. The depth of these edge domains decreases and the spacing between the edge domains increases with higher intensities. In Fig. 4(d), again domain-inverted features with widths of around 100 nm can be observed in the center of the stretched depth profile. Overall, the variety of domains shapes using a thin Cr coating in an ambient air are complex and somewhat surprising.
The unusual domain structures evident in Figs. 4(a) to 4(h) when a Z-cut LiNbO3 crystal, coated with 20 or 40 nm thin Cr layer was UV irradiated in an ambient atmosphere warrant further discussion. To guide this discussion, we attempt to identify the physical differences that would be expected for different Cr coating thicknesses:
First, the reflectivity of the Cr layer is assumed to be identical for the three coatings, and thus the heat deposition inside the crystal can likewise be assumed to be similar. The higher thermal conductivity of a thicker Cr coating is neglected. Second, the reduction of the crystal should depend on the Cr layer thickness. Under ambient conditions (air), oxidization of the Cr layer takes place with oxygen diffusing from both the air-Cr interface and the Cr-LiNbO3 interface, see also Fig. 2(b). This means that the thinner the Cr layer, the faster the Cr layer is saturated with oxygen from the surrounding atmosphere and therefore the crystal undergoes less reduction with a thin Cr layer than for a thicker Cr layer. As a consequence, the degree of reduction of the crystal becomes larger for thicker Cr layers at higher intensities. Third, when the Cr layer at the surface is completely oxidized, the driving force of the oxygen out-diffusion ceases and the domain inversion process caused by this process also ceases. Finally, the UV light intensity entering the crystal directly depends on the Cr layer thickness. The absorption length of Cr2O3 is estimated to be 30 times larger compared to a Cr layer (extrapolating the data from ). Therefore, more UV photons enter the LiNbO3 crystal when the Cr layer is completely oxidized. This oxidation occurs more quickly for thinner Cr layers according to the above-mentioned argument. The photon energy of the UV photons (5.1 eV) is greater than the band gap of LiNbO3 (~4 eV) and therefore electron-hole pairs are excited. Although most of the electron-hole pairs directly recombine and release their energy as heat, some might survive and drift under the influence of an internal electric field into the crystal. As a consequence, depending on both, the Cr layer thickness and the laser intensity, free charge carriers are generated within the top surface of the crystal.
The results from Fig. 4 together with the considerations above leads us to propose the following explanation for the domain inversion process for Cr coated LiNbO3 crystals in an ambient air atmosphere: The whole Cr layer is oxidized faster for thinner Cr coatings and higher laser intensities. Therefore, the larger absorption length of the generated Cr2O3 layer allows UV light to enter the crystal in these cases, where UV photo-induced electron-hole pairs are excited. The electrons will be highly mobile, whereas the positive ions will be less mobile. The region in which these electron-hole pairs are generated will follow approximately a Gaussian function laterally, corresponding to the profile of the incident UV beam that generates them. The electrons and holes will drift as a result of the Coulomb interaction along the diffusion gradient-based electric field and will therefore reduce this field. Additionally, the driving force of the oxygen out-diffusion is depleted, reducing the lithium in-diffusion and the diffusion-based electric field even more. The interplay between the diffusion gradient-based electric field and the UV induced free charge carriers, which hinder the poling, therefore defines the resulting domain shape. If the Cr layer is not completely oxidized, UV light will not reach the crystal surface and no free charge carriers will be generated. The electric field which is based on the diffusion gradient will simply be the inverse of the oxygen reduced area, which would lead to a nicely formed domain as it is illustrated in Fig. 5 (a) and was observed in Fig. 4(e). If the Cr layer is strongly oxidized, some UV light can enter the crystal and a small amount of free charge carriers would be generated. The number of charge carriers would only be sufficient to reduce the diffusion gradient-based electric field at the very surface. Therefore a ‘hollow’ domain emerges, corresponding to the illustration in Fig. 5(b) and the experimental observations in Fig. 4(f). If the Cr layer is completely oxidized even more UV light enters the crystal and a large amount of mobile charge carriers is generated which can drift deeper into the crystal, additionally the diffusion-based electric field is ceased due to the depletion of the oxygen sink at the surface. The charge carriers reduce the diffusion gradient-based electric field even more and domain inversion at the center is no longer possible. However, since the UV beam diameter is only 7 µm and the inverted domain width is ~10-12 µm (see Fig. 4), domain inversion still takes place in the area just outside the region that is directly illuminated with UV light. This is illustrated in Fig. 5(c) and experimentally confirmed in Figs. 4(g) and 4(h).
The very shallow domain-inverted region evident just beneath the illuminated region in Figs. 4(a)-4(d), 4(g) and 4(h) could be generated by a similar process like the standard UV domain writing on uncoated crystals, since, as is evident from Figs. 4(d) and 4(h) crossing domain-inverted features with widths of around 100 nm are observed, which are similar to those observed on the uncoated crystal in Fig. 3(b). The white stripes, which can be observed at the center of the irradiated tracks, could be a compound formed from oxidized Cr or an alloy formed by the Cr oxide and the melted LiNbO3 at the peak of the intensity distribution, which wasn’t stripped off by the Cr etchant.
3.4 Domain formation in 40 nm-thick Cr-coated Z-cut LiNbO3 crystals for different laser intensities in nitrogen atmosphere
In Section 3.3 it has been shown that the oxidization state of the Cr coating is critical for the generated domain shape. This suggests that a Cr layer that is only partially oxidized is necessary to achieve ‘Gaussian’ shaped domains. In Section 3.1 it was argued that using a nitrogen atmosphere restricts the source of oxygen to just lithium niobate, which should help to achieve full, solid domains also with thinner Cr coatings.
In order to test this hypothesis, a set of laser tracks with intensities between 2.14 × 105 and 2.44 × 105 W/cm2 were written on 40 nm Cr coated Z-cut LiNbO3 crystal within dry nitrogen atmosphere. The laser tracks were written with a line spacing of 30 µm. In accordance with the results shown in Fig. 1, the laser-irradiated tracks could be easily observed by the naked eye, as they appeared as dark stripes. Figure 6 presents SEM images of the UV written domains after wedge polishing and HF etching. As expected, the dry nitrogen atmosphere enables the formation of full, solid domains, with no evidence of hollowing. Similar results where also found for a Cr coating thickness of 20 nm only.
3.5 UV direct domain writing on a Cr-coated Y-cut LiNbO3 crystal
In Section 3.2 it was shown that it is possible to invert domains for both uncoated and a Cr coated Z-cut LiNbO3 crystals and it was concluded that the domain inversion mechanism in the two cases were due to two different physical processes. The domain inversion mechanism in the Cr coated crystal has the advantage of reduced surface damage when compared to the uncoated crystal. In the introduction it was mentioned that the UV direct writing process on uncoated crystals was able to generate domains on cuts with non-polar faces. Therefore, it would be advantageous to investigate whether the application of UV irradiation to the non-polar faces of Cr coated crystals will also result in domain inversion. An experiment was performed irradiating a 40 nm-thick Cr coated Y-cut crystal in a nitrogen atmosphere with UV laser light intensity of 2.14 × 105 to 2.44 × 105 W/cm2. The irradiation was performed with the focused UV spot being traced along the crystal surface in alternating directions. In accordance with the results shown in Fig. 1, the laser-irradiated tracks could be easily observed, as they appeared darker than the not irradiated crystal. Figure 7 presents SEM images of the cross-sections of the UV direct-written domains, after HF etching. It should be noted that these have not been wedge-polished as it was the case in Fig. 2, because the crystal Z-axis, which exhibits domain selective etching is already normal to the plane of the cross-section for this crystal cut. Figure 5 shows that domain inversion has taken place and that the width and depth of the inverted domain increased with increasing intensities. Furthermore, it can be observed that for lower intensities, Figs. 7 (a) and (b), the edge of the inverted domain is not well defined, but when higher intensities were used, Figs. 7 (c) and (d), this definition becomes stronger. Evidently, the domain inversion mechanism of oxygen out-diffusion by using a Cr coating offers similar flexibility to UV direct domain writing on uncoated LiNbO3 crystals.
In our previous studies of UV domain engineering on X- and Y-cut LiNbO3, it was found that the domain formation depended strongly on the direction in which the UV spot was scanned . We scanned the UV laser beam (2.73 × 105 W/cm2) along the + and -Z direction on a 40 nm Cr coated and an uncoated Y-cut crystal to test if also on Cr coated crystals such a dependence of domain formation on the scanning direction can be observed. Figure 8 presents the HF etched cross sections of the different scanning directions for a Cr coated and an uncoated crystal. On the Cr coated crystal no dependence of domain formation on the scanning direction of the UV laser beam can be observed, whereas for the uncoated crystal it is only possible to write a domain when the scanning direction is along the -Z direction. In Fig. 8(c) an inverted domain line can be seen directly under the UV written domain. The domain line was present occasionally at different depths, if a Cr coating was used. However, the exact conditions under which this domain line could be expected have not yet been identified. The fact that the domain formation does not depend on the scanning direction supports further more that the domain inversion mechanism is caused by a diffusion process, due to the isotropy of the diffusion.
3.6 Direct domain writing on Cr-coated 128° YX-cut LiNbO3
Up to now it has been shown that domain inversion is achieved by coating Z- and Y-cut LiNbO3 crystals with a thin Cr coating and irradiating tracks with a focused UV laser beam in a nitrogen atmosphere. The domain inversion mechanism has been attributed to a process that is triggered by the out-diffusion of oxygen, which reduces the crystal and results in a dark color of the UV-irradiated tracks. However, this dark color makes the domain engineering mechanism by oxygen out-diffusion less attractive for optical applications in the visible wavelength regime due to the high optical absorption. However, for SAW devices, the optical properties of the inverted domains are not critical and the similar SAW properties should be expected for reduced and non-reduced LiNbO3 [34,35]. The preferred crystal cut for SAW applications in the field of microfluidics is 128° YX-cut LiNbO3 . Thus it would be of interest to explore the effectiveness of the domain inversion process on this cut.
Direct domain writing was performed on a 128° YX-cut LiNbO3 crystal, which was coated with a 40 nm Cr layer. The domain writing was performed in a dry nitrogen atmosphere, with an intensity of 2.53 × 105 W/cm2 and a spacing of 21 µm between the scan lines. Figure 9 presents an SEM image of the periodically poled and wedge polished 128° YX-cut LiNbO3 crystal. Slight surface damage can be observed at the center of the written domains. The domain depth profile is similar to that observed for Z-cut crystals (Fig. 3). This shows that the UV direct writing with a Cr coating is also possible for LiNbO3 crystals with an inclined polarization axis such as 128° YX-cut LiNbO3. This demonstration represents an important new technology for the creation of acoustic superlattice structures  and an initial demonstration of the capabilities of this technique can be found in Yudistira et al. .
The fact that the inverted domains have no apparent dependence on the crystal orientation further supports the theory that the domain inversion mechanism is diffusion based.
We have shown that localized UV irradiation of a LiNbO3 crystal, which is coated with a thin film of Cr, results in localized domain inversion of the crystal. Evidence has been presented which suggests that the reactive metal coating (Cr) on the surface oxidizes at elevated temperatures and acts as an oxygen sink, drawing oxygen out of the crystal and reducing the crystal in the region beneath it. We have also shown that this domain inversion mechanism induces less thermal damage at the surface, compared to the standard UV domain writing process on uncoated LiNbO3 crystals. Additionally, it was found that this domain inversion mechanism works also at higher UV laser light intensities (2.44 × 105 W/cm2), if the UV writing is performed in a nitrogen atmosphere or when thicker Cr coatings are used. We finally presented a proof of principle for the technologically interesting non-polar crystal cuts, namely Y-cut and 128° YX-cut LiNbO3 for potential SAW applications. While a convincing demonstration of this technique has been presented and significant insight into this phenomenon has been gained, further investigation must be conducted in order to be conclusive about the underpinning physical phenomena that cause this behavior.
The authors acknowledge the facilities, and the scientific and technical assistance, of the Australian Microscopy & Microanalysis Research Facility at RMIT University and the Faculty of Engineering and Industrial Sciences at Swinburne University of Technology. We also want to thank Scott A. Wade for the fruitful discussions and the support in preparing the experiments. This work was made possible through the generous support of the ARC Centre of Excellence CUDOS. J.F. acknowledges the Vice-Chancellors Senior Research Fellowship from RMIT University and ARC DP1200013.
References and links
1. L. E. Myers and W. R. Bosenberg, “Periodically poled lithium niobate and quasi-phase-matched optical parametric oscillators,” Quantum Electronics, IEEE Journal of 33(10), 1663–1672 (1997). [CrossRef]
2. K. R. Parameswaran, R. K. Route, J. R. Kurz, R. V. Roussev, M. M. Fejer, and M. Fujimura, “Highly efficient second-harmonic generation in buried waveguides formed by annealed and reverse proton exchange in periodically poled lithium niobate,” Opt. Lett. 27(3), 179–181 (2002). [CrossRef] [PubMed]
3. M. Yamada, N. Nada, M. Saitoh, and K. Watanabe, “First-order quasi-phase matched LiNbO3 waveguide periodically poled by applying an external field for efficient blue second-harmonic generation,” Appl. Phys. Lett. 62(5), 435–436 (1993). [CrossRef]
4. J. Webjörn, V. Pruneri, P. S. J. Russell, J. Barr, and D. C. Hanna, “Quasi-phase-matched blue light generation in bulk lithium niobate, electrically poled via periodic liquid electrodes,” Electron. Lett. 30(11), 894–895 (1994). [CrossRef]
5. V. Gopalan, T. E. Mitchell, Y. Furukawa, and K. Kitamura, “The role of nonstoichiometry in 180° domain switching of LiNbO3 crystals,” Appl. Phys. Lett. 72(16), 1981–1983 (1998). [CrossRef]
6. M. Yamada and M. Saitoh, “Fabrication of a periodically poled laminar domain structure with a pitch of a few micrometers by applying an external electric field,” J. Appl. Phys. 84(4), 2199 (1998). [CrossRef]
7. F. Généreux, G. Baldenberger, B. Bourliaguet, and R. Vallee, “Deep periodic domain inversions in x-cut LiNbO3 and its use for second harmonic generation near 1.5 μm,” Appl. Phys. Lett. 91(23), 231112 (2007). [CrossRef]
8. S. Sonoda, I. Tsuruma, and M. Hatori, “Second harmonic generation in a domain-inverted MgO-doped LiNbO3 waveguide by using a polarization axis inclined substrate,” Appl. Phys. Lett. 71(21), 3048 (1997). [CrossRef]
9. E. L. Wooten, K. M. Kissa, A. Yi-Yan, E. J. Murphy, D. A. Lafaw, P. F. Hallemeier, D. Maack, D. V. Attanasio, D. J. Fritz, G. J. McBrien, and D. E. Bossi, “A review of lithium niobate modulators for fiber-optic communications systems,” Selected Topics in Quantum Electronics, IEEE Journal of 6(1), 69–82 (2000). [CrossRef]
10. G. Kovacs, M. Anhorn, H. E. Engan, G. Visintini, and C. Ruppel, “Improved material constants for LiNbO3 and LiTaO3,” Proceedings Ultrasonics Symposium (IEEE, 1990), pp. 435–438.
11. K. Nakamura and H. Shimizu, “Local domain inversion in ferroelectric crystals and its application to piezoelectric devices,” Proceedings Ultrasonics Symposium (IEEE, 1989), pp. 309–318. [CrossRef]
12. S. Miyazawa, “Ferroelectric domain inversion in Ti-diffused LiNbO3 optical waveguide,” J. Appl. Phys. 50(7), 4599–4603 (1979). [CrossRef]
13. H. Steigerwald, Y. J. Ying, R. W. Eason, K. Buse, S. Mailis, and E. Soergel, “Direct writing of ferroelectric domains on the x- and y-faces of lithium niobate using a continuous wave ultraviolet laser,” Appl. Phys. Lett. 98(6), 062902 (2011). [CrossRef]
14. A. C. Muir, C. L. Sones, S. Mailis, R. W. Eason, T. Jungk, A. Hoffman, and E. Soergel, “Direct-writing of inverted domains in lithium niobate using a continuous wave ultra violet laser,” Opt. Express 16(4), 2336–2350 (2008). [CrossRef] [PubMed]
15. A. Boes, H. Steigerwald, T. Crasto, S. A. Wade, T. Limboeck, E. Soergel, and A. Mitchell, “Tailor-made domain structures on the x- and y-face of lithium niobate crystals,” Appl. Phys. B 10.1007/s00340-013-5639-3 (2013).
16. A. Boes, T. Crasto, H. Steigerwald, S. Wade, J. Frohnhaus, E. Soergel, and A. Mitchell, “Direct writing of ferroelectric domains on strontium barium niobate crystals using focused ultraviolet laser light,” Appl. Phys. Lett. 103(14), 142904 (2013). [CrossRef]
17. P. B. Johnson and R. W. Christy, “Optical constants of transition metals: Ti, V, Cr, Mn, Fe, Co, Ni, and Pd,” Phys. Rev. B 9(12), 5056–5070 (1974). [CrossRef]
18. A. M. Mamedov, “Optical properties (VUV region) of LiNbO3,” Opt. Spectrosc. 56, 645–649 (1984).
19. C. L. Sones, S. Mailis, W. S. Brocklesby, R. W. Eason, and J. R. Owen, “Differential etch rates in z-cut LiNbO3 for variable HF/HNO3 concentrations,” J. Mater. Chem. 12(2), 295–298 (2002). [CrossRef]
20. M. Julkarnain, J. Hossain, K. S. Sharif, and K. A. Khan, “Optical properties of thermally evaporated Cr2O3 thin films,” Canadian Journal on Chemical Engineering & Technology , 3(4), 81–85 (2012).
21. S. Hallström, M. Halvarsson, L. Höglund, T. Jonsson, and J. Ågren, “High temperature oxidation of chromium: Kinetic modeling and microstructural investigation,” Solid State Ion. 240, 41–50 (2013). [CrossRef]
22. A. Dhar and A. Mansingh, “Optical properties of reduced lithium niobate single crystals,” J. Appl. Phys. 68(11), 5804–5809 (1990). [CrossRef]
23. K. L. Sweeney and L. E. Halliburton, “Oxygen vacancies in lithium niobate,” Appl. Phys. Lett. 43(4), 336–338 (1983). [CrossRef]
24. P. J. Jorgensen and R. W. Bartlett, “High temperature transport processes in lithium niobate,” J. Phys. Chem. Solids 30(12), 2639–2648 (1969). [CrossRef]
25. V. D. Kugel and G. Rosenman, “Domain inversion in heat-treated LiNbO3 crystals,” Appl. Phys. Lett. 62(23), 2902 (1993). [CrossRef]
26. L. Huang and N. A. F. Jaeger, “Discussion of domain inversion in LiNbO3,” Appl. Phys. Lett. 65(14), 1763–1765 (1994). [CrossRef]
27. G. Rosenman, V. D. Kugel, and D. Shur, “Diffusion-induced domain inversion in ferroelectrics,” Ferroelectrics 172(1), 7–18 (1995). [CrossRef]
28. K. Nakamura, H. Ando, and H. Shimizu, “Ferroelectric domain inversion caused in LiNbO3 plates by heat treatment,” Appl. Phys. Lett. 50(20), 1413 (1987). [CrossRef]
29. H. Steigerwald, M. Lilienblum, F. von Cube, Y. Ying, R. Eason, S. Mailis, B. Sturman, E. Soergel, and K. Buse, “Origin of UV-induced poling inhibition in lithium niobate crystals,” Phys. Rev. B 82(21), 214105 (2010). [CrossRef]
30. H. J. Donnerberg, S. M. Tomlinson, and C. Catlow, “Defects in LiNbO3—II. Computer simulation,” J. Phys. Chem. Solids 52(1), 201–210 (1991). [CrossRef]
31. J. M. Almeida, G. Boyle, A. P. Leite, R. M. De La Rue, C. N. Ironside, F. Caccavale, P. Chakraborty, and I. Mansour, “Chromium diffusion in lithium niobate for active optical waveguides,” J. Appl. Phys. 78(4), 2193 (1995). [CrossRef]
32. A. Dhar and A. Mansingh, “On the correlation between optical and electrical properties in reduced lithium niobate crystals,” J. Phys. D Appl. Phys. 24(9), 1644–1648 (1991). [CrossRef]
33. V. Y. Shur, D. K. Kuznetsov, E. A. Mingaliev, E. M. Yakunina, A. I. Lobov, and A. V. Ievlev, “In situ investigation of formation of self-assembled nanodomain structure in lithium niobate after pulse laser irradiation,” Appl. Phys. Lett. 99(8), 082901 (2011). [CrossRef]
34. P. F. Bordui, D. H. Jundt, E. M. Standifer, R. G. Norwood, R. L. Sawin, and J. D. Galipeau, “Chemically reduced lithium niobate single crystals: Processing, properties and improved surface acoustic wave device fabrication and performance,” J. Appl. Phys. 85(7), 3766 (1999). [CrossRef]
35. E. M. Standifer, D. H. Jundt, R. G. Norwood, and P. F. Bordui, “Chemically reduced lithium niobate single crystals: processing, properties and improvements in SAW device fabrication and performance,” Proceedings of the 1998 IEEE International Frequency Control Symposium (IEEE, 1998), 470–472. [CrossRef]
36. J. Friend and L. Y. Yeo, “Microscale acoustofluidics: Microfluidics driven via acoustics and ultrasonics,” Rev. Mod. Phys. 83(2), 647–704 (2011). [CrossRef]
37. D. Yudistira, S. Benchabane, D. Janner, and V. Pruneri, “Surface acoustic wave generation in ZX-cut LiNbO3 superlattices using coplanar electrodes,” Appl. Phys. Lett. 95(5), 052901 (2009). [CrossRef]
38. D. Yudistira, A. Boes, A. Rezk, L. Yeo, J. Friend, and A. Mitchell, “UV direct write Metal Enhanced Redox (MER) domain patterning for writing surface acoustic piezoelectric superlattice in lithium niobate,” Adv. Mater. Interfaces. submitted.