Photoinduced microdomains

Photomicrodomain scattering indicatrix is well enough described within the approach of scattering by long thin cylinders (needles). What formation can play role of such needles? This could be antiparallel domains. Indeed, such microdomains can be formed in lithium niobate crystals due to the symmetry properties. However repolarization of lithium niobate crystals is known to be hampered for large values of both the Curie temperature (T c=1480 o K) and coercive field. Pure lithium niobate used to be considered to be so called frozen ferroelectric. Its coercive field was estimated to reach the value about 15 million V/cm. Thus repolarization of such crystal used to look impossible due to electric destroying the sample. Defects of the crystal structure, admixtures etc. diminish the coercive field value. But it still remains very high and reaches the value about E c ~ 300 kV/cm. So repolarization of lithium niobate is quite hardly realized but possible. Besides the process of lithium niobate crystal repolarization does start by appearance of needle-like microdomains. Such microdomain structure could be displayed by etching the crystal surfaces in acid mixture (hydrochloric acid, sulphuric acid, hydrofluoric acid). The picture demonstrates an initial stage of lithium niobate crystal repolarization under external electric field about E ext= 85 kV/cm and temperature T= 130 oC. Our experiments as well displayed the microdomain structure having appeared after photoexciting the crystal.

Investigating an angle dependence of spatial distribution of the scattering radiation allowed us to calculate average dimensions and surface density of the domains: length was about 1-2 mm, diameter 1 µ, density 10 6 cm -2 .

So a set of antiparallel (relative to polar axis) needle-like domains originates under light influence in our crystal. Does some electric field playing role of external depolarizing one exist in the crystal? Of course it does. This is a field of redistributed photoinduced charge (the last takes place due to photovoltaic effect). By our measurements this field reaches the value about 70 kV/cm. Unfortunately, photovoltaic current flows toward polar axis so that the resulting photoinduced field has a direction inappropriate to repolarize the crystal. Moreover, it only could strengthen singledomainity of the sample. What can help us to understand how antiparallel microdomains appear under photoexcitation of the crystal?

Let us remember that the conic scattering instantly disappears after placing the crystal into conductive media. Hence the role of surface charges is considered to be important for domain forming. Direct measurement of integral surface charge value Q brought such a result: Q = 1.7 10 -10 Coulombs independently on PL at P ≥ 10mW. This is quite a large charge.

The effect of unlocality also denotes a possibility of important role of surface charge state because it is surface charge carriers to be able to move far enough from illumination region and cause there some appropriate conditions for repolarization.

Investigation of surface charge density distribution σ(x,y) was performed by electrographic development method. This method uses some charged powder dispersed in nonconductive liquid (in our case chlorine carbide CCL 4). As a result of such electrographic development dark powder adheres to positively charged part of surface.

 

At -z surface the positive charge has been found only to concentrate in the pump localization area. At the same time at opposite “+z” surface, the negative charge appeared to be localized at the centre of the pump illumination area making a circle with the diameter of 4 - 5 mm. Besides the positive charge formed a ring - shaped zone with average diameter of ~12 mm. It is in that zone that the conic scattering of testing beams (and consequently, microdomains) appeared first of all.

The results described above can be explained within the supposition that the photovoltaic effect  is the main cause of the charge generation at z-surfaces of the crystal. Direction of photoinduced electron motion due to the photovoltaic effect (from “-z” side to “+z” side) confirms the supposition. Besides at “-z” surface, positively charged deep traps (Fe 3+ ions) remain motionless (immovable) and only localized in the area illuminated by the pump. At “+z” surface, mobile photoinduced electrons having filled up all the deep traps, begin shifting under influence of the Coulomb forces and finally form a circle with the diameter to be bigger than exitant beam one. At the same time the positive charge carriers (for instance, oxygen vacancies or surplus ions Li  +) begin flowing from the crystal edge to the negative charge localization area making the positive charged ring pointed out above. At that ring-shaped area, 180  - repolarization is energetically profitable as negatively charged butt-ends of the microdomains reduce both the surface charge density and surface energy density U~. At central negatively charged circle-shaped zone 180 o - repolarization only could increase the surface charge density. However 180 o - microdomain forming at that zone is profitable from another point of view : it brings to decrease of photovoltaic current value since repolarized regions generate the photovoltaic current in invert direction. Of course, such repolarization takes place at less favourable conditions compared with positively charged ring-shaped zone.
Note that optically undistinguishable (under usual conditions ) 180 o domains can be distinguishable under influence of the photovoltaic electrostatic field Eg due to electrooptical effect. Besides, accordingly to the surface charge distribution, the field Eg(r) has significant transversal components E gx,y that brings to the rotation of the passing light polarization plane.The appropriate angle of turning of polarization plane reaches the value about /2 in the thick crystals (thicker than 1mm) making the conic scattering practically independent on initial light polarization outside of the crystal.