AmericanChemicalSociety.com

The main objective of the project was to develop a general theory of conductivity in oxides with wide band-gap (3 eV and above), namely, the oxides of main group metals such as post-transition metals Zn, Ga, In, Cd, Sn as well as light metals Mg, Ca, Al, Si. All these oxides share the same electronic configuration of cations (namely, s0), yet, their electrical properties are very different ranging from classic insulators to nearly-metallic conductors. Our goal was to understand the origin of electron localization and to find ways to achieve a stable conductive state via efficient free-carrier generation and, at the same time, to maintain a sufficiently low optical absorption within the visible.

Our comparative electronic band structure investigations of main group metal oxides have led to several important conclusions which are briefly summarized below.

1) The unique electronic and optical properties of typical transparent conductive oxides (TCO) originate from the strong interaction between the metal s-orbitals and the p-orbitals of the neighboring oxygen atoms. Due to the Ms-Op overlap, the undoped stoichiometric oxides exhibit large band gaps (3.4-4.9 eV) and small electron effective masses (0.28-0.35 me) which cannot explain the differences in the observed electrical conductivities when the materials are degenerately doped.

2) Little sensitivity of the Ms-Op overlap and, hence, of the electron effective mass to structural variations - including variations in the oxygen coordination, strong distortions in the polyhedra, irregular atomic arrangements and large structural voids - explain the success of (i) amorphous TCOs whose optical and electrical properties remain similar to those in the crystalline state, and (ii) TCO-based flexible transparent conducting coatings.

3) Our comparative studies of post-transition and light-metal wide-bandgap oxides have revealed that the energy location of the cation(s) empty p-states with respect to the conduction band bottom plays the key role in determining the transport properties in these materials. Our results explain the trend in the observed conductivities/carrier mobilities in the conventional TCOs such as In2O3, ZnO and Ga2O3, and explain why light-metal oxides (Al2O3, CaO, MgO) remain classical insulators.

4) Most significantly, based on our findings, we suggested several ways to overcome the detrimental electron localization effects by tuning the resulting electronic properties via (i) chemical composition (multicomponent oxides); (ii) unusual local structure (nano-porous Ca12Al14O33); or (iii) specific carrier generation.

The efforts of the PI on this project were three-fold:

1. To advance current understanding of the underlying physics in conventional transparent conductive oxides via a thorough, comparative and accurate studies of most-known/most-employed TCO materials (In2O3, ZnO, CdO, Ga2O3, SnO2, and InGaZnO4) that allows for (i) prediction of ways to controllably manipulate the electronic and optical properties; and for (ii) continued search of novel efficient transparent conductive materials;

2. To educate young researchers on the topic and to guide them through the processes of learning and mastering the computational electronic-band structure techniques; and

3. To disseminate the knowledge gained from the studies via publications, conferences/meetings participation, and collaborations with other researches at academia and industry.

Dr. Min Sik Park (postdoctoral fellow in the PI's group) has performed first-principles calculations of transport coefficients for several oxides. This recently-developed capability was implemented within highly precise all-electron full-potential linearized augmented plane wave method (FLAPW). This code allows us to calculate temperature- and carrier-concentration-dependent electrical conductivity (sigma) and Seebeck coefficient (S) based on the distribution function given by Boltzmann equation in the constant relaxation time approximation. For the calculation of the group velocity which is included in the transport coefficients, we use full intra-band optical matrix elements defined within FLAPW method.

In particular, Dr. Park has performed investigations of the electronic, optical and thermoelectric properties of highly reduced Ca12Al14O33 (also known as mayenite) were investigated. The results of our detailed band structure investigations provide an insight into the observed insulator to metal transition in cage-structured mayenite. Moreover, based on the peculiar electronic band structure of the oxygen deficient Ca12Al14O33 , namely, the presence of a soft Coulomb gap near the Fermi level, we explain the observed change in the sign of the Seebeck coefficient. Our calculated temperature and carrier concentration dependent Seebeck coefficient and electrical conductivity are in excellent agreement with the experimentally observed ones.

Further, in search for novel thermoelectric oxide materials, Dr. Park has calculated the temperature- dependent Seebeck coefficient S and electrical conductivity in layered multicomponent InGaZnO4 substitutionally doped with Sn, Ti, V and the one with an oxygen vacancy. To improve thermoelectric transport properties (and, in particular, S) in InGaZnO4 , we considered substitution with 3d transition metals Ti and V for In atoms in InGaZnO4. Indeed, Ti and V 3d states induce narrow bands at 1.3 and 0.8 eV above CBM, respectively. These narrow bands lead to the DOS and group velocity both having a large slope at the Fermi level. As a result, Ti- and V-doped InGaZnO4 show enhanced S below 300K as compared to the stoichiometric case. By using the power factor (PF) of 6.37x10-4 Wm-1 K-2 and the thermal conductivity of 5 Wm-1 K-1 at 300K, the largest ZT of 0.38 is obtained in the V-doped case. In the Ti case we obtain ZT=0.31. Both cases result in significantly higher ZT compared to the experimentally observed one for InGaZnO4 (ZT=0.05) which, as expected, is similar to our calculated figure of merit for Sn and VO (ZT=0.02-0.03).

Dr. Min Sik Park, who joined the group in October 2008, has been 90% supported by this project.

The results of this project have been either published or prepared for publication: one book chapter by the PI is in production, 1 paper has been accepted to Physical Review B, and 3 manuscripts are ready to be submitted for specific journals. In addition, another 2 manuscripts are in a draft stage of preparation for publication. All necessary calculations have been finished for these manuscripts, the results have been analyzed and prepared for publication in the form of tables/figures, and the first drafts have been prepared.