First-principles study of LiNbxM1-x O3 (M= V, W, Ta, Mo,…) for holographic memory applications

Abstract: For holographic memory applications, the photorefraction of well-known ferroelectric such as lithium niobate doped with different transition metals is very important. First principles study assumes special significance in this context, as to why certain transition metal atoms are better than the other. In this work, Nb atom in LiNbO3 was substituted with transition elements having valency greater than/equal to +5 for UV photorefraction applications, and atomistic first-principles calculations were done using HSE06 functionals. The d-states of the transitional elements were found to decrease the band-gap of the host material having implications for a suitable material design. Minimum band-gap was obtained for W, while Ta showed a maximum value. Absorption coefficients were estimated for each material and based on their low values at 351 nm (i.e. for holographic applications) that is the usual UV photo-refraction wavelength, the elements found suitable were V, W, Ta, Mo. Then birefringence properties for these crystals were also studied to predict that V and W were good candidates.

Some Descriptions: Lithium Niobate (LN) has gone through intensive research for its volume holographic data storage capability (Ashkin et al., App. Phys. Lett. 1966. 9, 72). The theory of the volume phase holograms in the birefringent electro-optic crystals  (LN) has been developed a long time ago via electrically controlled light diffraction and pattern reconstruction at a "READ" wavelength different from that of the "WRITING" beams (Petrov et al, Ferroelectrics 1978, 21, 631).

        Holographic data storage is a technique that can store information at high density inside the crystals with the volume of the recording media used for data storage instead of just the surface. Opening up the inside of a material for storage, rather than just using its surface, yields huge improvements in capacity. In addition, unlike conventional technologies, which record and recover digital data bit by bit, holography has the ability to "read and write" millions of bits of data with a single pulse of light, enabling data-transfer rates of billions of bits per second with storing capacity of several Tbs.

            This subject remained dormant for quite some time. After a long break, the subject has been revived by an intensive activity on LN crystals doped with different transition elements (McMillen et al, Opt. Exp. 1998, 2, 491). The large crystal size of LN allows: (a) high capacity, (b) angular selectivity, (c) long storage lifetime, and (d) high diffraction efficiency. LN can also be used in optical "correlator type holographic data storage" applications, where analog images are accessed at high speed. In particular, angle-multiplexing type storage in LN has been observed where ‘optical correlation’ is used for data storage (Mok et al., Opt. Lett. 1993, 18, 915).

      Transition metal doped LN ferroelectric material is used for photo-refraction and volume data-storage. Although a multitude of such materials have been characterized in details, the explanation as to why only a few of them are preferred compared to others is still to be known. In the field of condensed matter physics, one of the well-studied materials is ferroelectric, which has a variety of important applications in nonlinear optics, such as electro-optics and second harmonic generation, and non-volatile memory devices (see Bandyopadhyay et al, Phys. Rev. 2010, B81, 064104, and references therein). Ferroelectrics have also emerged as important materials as piezoelectric transducers, pyroelectric detectors, surface acoustic wave (SAW) devices, four-phase mixing doublers. Such applications and many more as 'nano devices' are mentioned in the latter reference given above. One of the most important ferroelectric materials is LN that is widely used as an optical material for its various high frequency photonic and optoelectronics applications.

      Commonly Fe-doped LN is used for writing holograms in blue-green region of the visible spectrum, but in other part of electromagnetic spectrum, the ‘sensitivity’ is needed for inexpensive applications. In addition, LN:Fe has problems, such as  low response speed, strong light-induced scattering and volatility. Other transition elements such as Cu, Mn, Ni etc. have also been investigated for holography. Early attempt of doping with more than one transition elements in LN was made for various systems. Among transition elements, which were found to be suitable for holographic memory applications are: Sc, Zr, Hf, V, Mo, Mn, Fe, Ru and Cu. It was also found that the substituents occupying Li-sites were those having valence less than 5+ Hence, others were found to occupy Nb sites and were primarily used for studying UV-photo-refraction.

      Despite intensive experimental work, comparatively less theoretical investigation has been carried out on this subject, as well as on density functional theory (DFT) approach for atomistic descriptions. One of the pioneering investigations on the electronic structures and energetics of LN in comparison with those of another ferroelectric (LT) was done by Inbar and Cohen (Phys. Rev. 1996, B53, 1193) via ‘linearized augmented plane-wave’ (LAPW) frozen phonon calculations in order to shed light on phase transition and origin of ferroelectricity. Now, for the electronic and optical properties of LN, various levels of DFT calculations such as general gradient approximation (GGA) and GW calculations were carried out, and the "band gap" of LN was a crucial topic in the last few decades

            It has been evident that there is a necessity of studying the atomistic electronic properties, particularly in relation to 'band-gap engineering' of such photorefractive materials involving LN via density functional theory (DFT) approach. This assumes more importance due to the recent surge in the activity in the field of holographic memory applications. In the present work, for "holographic memory" applications, the photo-refraction of well-known ferroelectric such as lithium niobate doped with different transition metals is very important. First principles study assumes special significance in this context.

            In this work, Nb atom in LiNbO₃ was substituted with transition elements having valency greater than/equal to +5 for UV photo-refraction applications, and atomistic first-principles calculations were done using HSE06 functionals. The d-states of the transitional elements were found to decrease the band-gap of the host material having implications for a suitable material design. Absorption coefficients were estimated for each material and based on its low values at 351 nm (for holographic applications) that is the usual UV photo-refraction wavelength, the elements found suitable were V, W, Ta, Mo. Then birefringence properties for these crystals were also studied to predict that V and W were good candidates. Here, we study the dopants that have been used (or possibly used) for photo-refraction having valence more than 5+, namely V, Ta, Cr, Mo, W, Tc, Re, Ru, Os, Rh, Ir, Pt by using HSE06 hybrid functional for these calculations.

Finally, in the realm of 'band-gap engineering', the results show that tantalum (Ta) has the minimum impact, while tungsten (W) has the maximum effect on the host crystal on the reduction of 'band-gap' within the systems under the present investigation. Then, it was interesting to study the density of states (DOS) of such materials both in spin-up and spin-down conditions. It is quite pertinent to mention about an important work by Tong et al. (Phys. Rev. 2014, B89, 064404), who observed significantly different optical responses from spin-up and spin-down channels in case of multiferroic EuO via first principles DFT calculations.

Moreover, among few elements found suitable in this work, Ta is very well-known as nonlinear optical material showing the lowest absorption coefficient for LN:Ta at 0.1, although its band-gap is the highest (5.07 eV). However, for LN:W whose absorption coefficient is also very low at 0.2, its band-gap is the lowest (0.72 eV). This opens up a new field of work in the realm of "Band-Gap Engineering" of device material formulations to suit different applications. For example, in a recent work by Grinberg et al (Nature 2016, 503, 509), a 10% replacement of KNbO₃ by nickel-niobium based material decreased the band-gap to 1.39 eV to give rise to a very high level of photovoltaic efficiency in terms of high "photocurrent density for solar-energy conversion" and other important applications for this type of ferroelectric-semiconductor based materials.

It is pertinent to mention that in the present investigation with our DFT approach, a comparable band-gap was observed for LN:Re and LN:Rh and LN;Mo at 1.30 eV, 1.36 eV and 1.57 eV respectively. These materials could also be considered as competitive candidates for photoconduction based "Solar Energy" applications.

http://arxiv.org/abs/1709.01970