METHODS, IDEAS, APPLICATIONS
My work covers a wide range of frontier research topics in Condensed Matter Physics and Materials Research. The research projects I have been involved with are extended to both metals and semiconductors; their main characteristic is that they require highly sophisticated computational techniques in order to be studied.
PROJECT CLASSIFICATION
My research projects can be classified as follows :The single-impurity problem has been studied at various levels of approximation using a variety of embedding methods that we have developed or extended (KKR, SSH, CHF, embedding etc, as described in paragraphs B.1.2, B.1.3 and B.1.4 below).
The many-impurities problem ( and the problem of alloys ) has been attacked by generalizing the Coherent Potential Approximation (CPA) method (as described in paragraphs B.1.5 and B.1.7) and by cluster approaches (as deSCribed in paragraphs B.1.3 and B.1.6 below).
NEW IDEAS - MAIN CONTRIBUTIONS
My major contributions include the following new theoretical methods that we proposed (in chronological order) :
All the proposed theories have been also used in developing new computational approaches; new state of the art computer codes have been developed. Concurrently with the developing of the new theoretical and computational approaches, I have also extended existing computational methods according to the requirements of my work (as explained below). My emphasis has been focused at explaining current experimental data and at introducing efficient computational methods of ab initio character or of a firm ab initio footing. I tried to develop computational techniques which have the potentiality to be general enough so as to be easily applicable for a wider class of problems than for the problem for which they were initially inspired and designed. In the following, I will restrict my presentation only to those contributions (new ideas and/or methods ), which, to my opinion, are the most significant ones among those that I have proposed in my published work.
DETAILED DESCRIPTION OF MAJOR CONTRIBUTIONS
Abstract : The Periodic Planar Jellium Approximation, (PPJ-A), that we have proposed, is a generalized jellium model for which the positive charge of the metal ions is assumed to be uniformly distributed along the lattice planes which are parallel to the metal surface. For Bravais lattices, tHe PPJ-A leads to a uniform distribution of the positive charge on equidistant planes parallel to the surface. Thus, within the PPJ-A, one retains explicitly the lattice periodicity along the direction perpendicular to the surface. As a result, the proposed model allows one to take explicitly into account the effect of surface anisotropy onto the electronic properties of the metal surface, i.e. work function, surface energy, etc., which are experimentally measured quantities. Furthermore, the proposed model has been proved suitable to simulate imperfect, chemisorbed and implanted metal surfaces. Recently, this model has been found more efficient in describing clusters of simple metal atoms as compared with the usual spherical jellium drop approximation [see J.Lerme et al, Phys. Rev. B50, 5558 (1994); J.Lerme, Phys. Rev. B54, 14158 (1996)].
Coupled Hartree Fock Approach to embedding
Abstract: The Coupled Hartree Fock Approximation, (CHF-A), is a non-local Hartree-Fock, (HF), method, originally proposed by Cohen and Roothaan [J. Chem. Phys., 43, 534 (1965)], for studying atoms in external electric and magnetic fields. We have generalized this method, at various levels of approximation, in order to treat atoms and clusters embedded in crystal fields either on the surface or in the bulk.
Within the CHF-A that we have proposed, one utilizes the potentiality of the chemical methods (non-local HF and beyond ) to study changes at the atomic/molecular level in atoms/clusters embedded in crystal fields, the latter being described at a poorer level of approximation (i.e. the LDA or the LDA-based PPJ-A and usual jellium approximation ). The method was found very efficient in describing chemisorpion problems as well as the the interaction of noble gas atoms with metal surfaces (i.e., He, Ar with Na(001) and Al(001) surfaces) the metals being described within the PPJ-A described in B.1.1. The method was also used to study the embedding of noble gas atoms in metallic jellium.
Subsequently, our calculational scheme of the CHF-A was combined with the embedding scheme of Inglesfield [ J. Phys. C: Solid State Phys. 14, 3795 (1981)] and the ideas of the Complex Coordinate Method, (CCM), [see Milfeld and Moiseyef, Chem. Phys. Lett. 130, 145 (1986); Attabek et al , Mol. Phys. 40, 1107 (1980) ], in an attempt to study excited atoms on metal surfaces. We demonstrated the applicability of our method by calculating the state of the excited Ar on Na(001) surface. Our proposed generalization, to be called `Inglefield's Complex Potenital, (ICP), CHF-A' and briefly as ICP-CHF-A , is a very promising approach for undertaking studies of atoms and clusters ( being either in their ground or excited state ) interacting with metal surfaces or being embedded in bulk crystals. (This method an alternative of the well known Surface Green's Function Matching (SGFM) method ), Such studies are of significant importance in the description of dynamical processes on surfaces (catalysis). The potentiality of our method is becoming more and more appreciated as it becomes clearer that accuracies achieved by the standard " chemical methods " are also required to understand and describe a vast number of experimental data of the current Condensed Matter Physics (see for example, Abarenkov et al , Phys. Rev. B56, 1743 (1997) and E. Carter et al, Phys. Rev. Lett. (2001) ).
Very recently we appied this theory (within Fisher's proposed modification of Inglesfield's method - see A.J.Fisher, J.Phys. Condens. Matter, 2, 6079 (1990)) in the study of transport properties of carbon nanotubes. For this application we made use of the Tight-Binding description of the (metal or semiconducting) host as described in paragraph B.1.6 below.
Various computer codes have been developed for the CHF-A and the ICP-CHF-A respectively. As from its nature, the method of the ICP-CHF-A solves a complex Hamiltonian problem contrary to the case of the CHF-A which solves a real Hamiltonian problem. Our applications consider the metal host either at the jellium approximation or at the LDA of the more realistic periodic lattice .
The Sub-Space Hamiltonian approach to the embedding of atoms and clusters in semiconductors : The reconstruction of the Si(110) surface and its interaction with Si and Ni adatoms
Ab-initio Calculations in Metals with Impurities within the Quadratic Approximation to the Korringa Kohn Rostoker Method
Our method introduces a fast and efficient calculational scheme which allows one to study impurity resonances, charge transfer and solution energies of metals. Our impurity code is currently used to develop a fast version of the Coherent Potential Approximation (CPA) within the QKKR approach, ( the method to be called QKKR-CPA ), for treating binary metal alloys.
Studies of Binary Alloys within the Coherent Potential Approximation
The Tight Binding Molecular Dynamics method in the Hubbard approximation - Applications
More recently, we combined our TBMD method with the embedding scheme ICP-CHF-A described in paragraph B.1.2 above and proposed a new theory for studying the transport properties of single wall carbon nanotubes (SWCN's). The theory was applied successsfully in the studies of the tunneling conductivity of both straight and Y-shaped SWCN's.
As the study of complex systems is far from the reach of present day ab initio computational methods, our method opens a new approach to study these systems using a semi- empirical method whose ab initio footing can be well controlled.
In this project, we studied the dependence of the optical absorption coefficient of the Si(m)Ge(n) SL's on the roughness of the Si-Ge interfaces, the latter approximated by a Si(x)Ge(1-x) substitutional binary alloy whose Si-concentration x depth profile simulates the degree of the interface roughness. We have shown that, in agreement with available experimental data, the optical absorption coefficient of the Si(n)Ge(m) SL's is degraded as the roughness is inceased. The degradation was found to depend on the Si depth profile along the Si-Ge interfaces.
Since the early investigations of magnetism in d0 materials, a consensus has emerged in attributing this magnetism to the presence of defects. Furthermore, the most recent and outstanding works on diluted magnetic semiconductors (DMS) showed that the magnetic properties are not exclusively related to the presence of the magnetic ions but strongly determined by the defects. These can be intrinsic (e.g., any structural and/or topological defects) or extrinsic (e.g., any impurity atoms or radicals) or a combination of them. Soon, the implementation and the management of defects were considered as powerful tools for inducing and tailoring magnetic features in classes of systems of great technological interest which are not magnetic otherwise. These systems include; DMS, transition metal oxides (TMOs), carbon based materials (graphene ribbons, irradiated carbon), ordinary oxides (e.g., CaO, HfO, etc).
The development of defect magnetism in these materials appears to be a two step process. During the first step, unpaired electrons are introduced which provide the magnetic moments (MMs); this is one of the roles that is attributed to the defects. In the second step the magnetic coupling (MC) among the MMs is developed. While the origin of the MMs (either intrinsic or extrinsic) is undoubtedly attributed to the presence of the defects, the origin of the MC among the MMs is still far from being well understood despite the plethora of reported experimental and theoretical investigations and the models proposed at various levels of computational approximations.
Defect states and formation of impurity bands induce dramatic alterations in the band structure of the wide band gap or semiconducting host materials and may turn them into metals. Unavoidably, in this case, the magnetism of metallic DMSs and doped TMOs was attributed to carrier mediated processes (Ruderman–Kittel–Kasuya–Yosida (RKKY), double exchange, p–d and s–d exchange) and the factors they affect them. On the other hand, in the case of non-metallic DMSs and doped TMOs, the corresponding MC between two magnetic dopants was attributed to the mediation of an anion or of a radical, the latter made of a codopant and its surrounding anions. The observation that magnetic features in non-metallic DMSs and doped TMOs are found even for dopant concentrations below the dopant percolation threshold led to the proposal of various models for its justification.
Our work introduces a new perspective in explaining the origin of magnetism in DMSs, TMOs, carbon-based materials and other related materials. According to our proposal, the magnetism in these materials is the result of synergistic action of defect-induced electronic processes mostly of local character which can provide magnetic moments and develop a ferromagnetic coupling among them. This synergy is realizable via appropriate codoping which appears as a general and generic approach.
Our research activity on Diluted Magnetic Semiconductors (DMSs) has been intrigued by our results obtained for the 2D magnetic Rhombohedral-C60 polymer according to which the ferromagnetism in this material can be attributed to the simultaneous presence of two kinds of defects one acting as a “donor” and the other as an “acceptor” type. As a result of this combination of defects, lone electron spins are produced while charge transfers are developed which lead to remote overlapping among the molecular orbitals and create electric fields which can sustain a ferromagnetic coupling between the lone spins.
The vacancy model that we proposed for the magnetic properties of the 2D C60-based polymers appears as a generic model for magnetism in systems with defects including in addition to the carbon-based materials, many, if not all, members of DMSs (as for example, ZnO, TiO2, CaO), III-V and II-VI semiconductors, hexaborides, etc. Guided by our results for the magnetic 2D C60-based polymers, we have proposed that the magnetism in all these materials can be developed and enhanced if they are appropriately codoped with a pair of impurities with at least one of them be able to provide lone spin electrons. This model has been applied with success in the case of TM-codoped systems based on ZnO, GaN, GaP, MoS2, TiO2, SnO2, CdS, ZnS etc. Our initial results for the ZnO “codoped” with Co and Cu atoms, within the DFT/SGGA+U approximation, did in fact justify our hypothesis as the Zn(Co,Cu)O system was found to exhibit a FM ground state while the singly doped Zn(Co)O was found to be AFM.
Codoping is the simultaneous doping of a material by two or more dopants. In a series of publications I have demonstrated that Codoping introduces a new concept in both the experimental and theoretical investigation of the defect magnetism; it appears as a set of cooperative processes undertaken by the codopants which have been labeled defect-induced defect-mediated (DIDM) model. This emphasizes the two step process for establishing defect magnetism. That is, the first contributes in the formation (intrinsic or extrinsic) of the defect-induced magnetic moments and the second participates in the establishment of the defect mediated magnetic coupling among the maagnetic moments.
Codoping brings into consideration the spin polarization induced by the magnetic moments on their anion ligands and empasizes their role in the establishment of the magnetism in DMSs and dTMOs.
Key factors for the establishment of the magnetic coupling (MC) within the defect-induced defect-mediated (DIDM) magnetism are attributed to:
We have shown, using ab initio DFT calculations, that the DIDM magnetism is associated with local and holistic responses of the defected host to the presence of defects. Codoping focuses mainly on the local aspects of the DIDM; this manifests itself within bipartite dopant conformations introducing new contributions to the MC, namely the successive spin polarization (SSP) and the successive superexchange (SSX) ones, which are competitive to well known classical ones (e.g., superexchange, double exchange, etc). The holistic aspect of the DIDM is manifested in the energy shifts of the system’s d-band and p-band centers which affect the energy band gap. (See TOPICAL REVIEW appearing as Ref.153 in the List of Publications ).
We have identified an exciting new discovery of a relationship between magnetism and band gap dynamics in the field of diluted magnetic semiconductors (DMSs) and doped transition metal oxides (dTMOs). This is the result of a systematic analysis of ab initio numerical results for several magnetic dopants in DMSs and dTMOs which points to an intriguing connection between the local and the holistic features of these systems. This relationship is found to be the outcome of the orbital hybridization, namely that of the dopant d-bands with those of the sp host anions and is interpreted as the inverse analogue of the well known adsorption process, that of sp-atoms onto metallic surfaces. We show that the dopant complexes act as large impurity atoms or megatoms which become hybridized by the host anions. As a result, the magnetic hybridization energy becomes describable in terms of well known adsorption model approximations (e.g., Newns, Norskov), (See Ref.154 in the List of Publications ).
The theory of the codoping that we are proposing can be used as a practical guide for choosing the appropriate host and pair of codopants for a specific semiconductor environment that can lead to the fabrication of DMSs and TMOs with enhanced magnetic properties.