1: % Template article for preprint document class `elsart'

2: % SP 2006/04/26

3: 

4: \documentclass{elsart3p}

5: 

6: % Use the option doublespacing or reviewcopy to obtain double line spacing

7: % \documentclass[doublespacing]{elsart}

8: 

9: % if you use PostScript figures in your article

10: % use the graphics package for simple commands

11: % \usepackage{graphics}

12: % or use the graphicx package for more complicated commands

13: \usepackage{graphicx}

14: % or use the epsfig package if you prefer to use the old commands

15: \usepackage{epsfig}

16: 

17: % The amssymb package provides various useful mathematical symbols

18: \usepackage{amssymb}

19: 

20: % The lineno packages adds line numbers. Start line numbering with

21: % \begin{linenumbers}, end it with \end{linenumbers}. Or switch it on

22: % for the whole article with \linenumbers.

23: % \usepackage{lineno}

24: 

25: % \linenumbers

26: \begin{document}

27: 

28: \begin{frontmatter}

29: 

30: % Title, authors and addresses

31: 

32: % use the thanksref command within \title, \author or \address for footnotes;

33: % use the corauthref command within \author for corresponding author footnotes;

34: % use the ead command for the email address,

35: % and the form \ead[url] for the home page:

36: % \title{Title\thanksref{label1}}

37: % \thanks[label1]{}

38: % \author{Name\corauthref{cor1}\thanksref{label2}}

39: % \ead{email address}

40: % \ead[url]{home page}

41: % \thanks[label2]{}

42: % \corauth[cor1]{}

43: % \address{Address\thanksref{label3}}

44: % \thanks[label3]{}

45: 

46: \title{The ground state phases  of orbitally degenerate spinel oxides}

47: 

48: % use optional labels to link authors explicitly to addresses:

49: % \author[label1,label2]{}

50: % \address[label1]{}

51: % \address[label2]{}

52: 

53: \author{George Jackeli\corauthref{cor1}}

54: 

55: \address{Institute for Theoretical Physics,

56: Ecole Polytechnique F\'ed\'erale de Lausanne, CH-1025, Lausanne,

57: Switzerland,\\ and E. Andronikashvili Institute of Physics, Tbilisi, Georgia.}

58: \corauth[cor1]{Tel: +41 21 693 44 79: fax:   +41 21 693 54 19.\\

59: {\it E-mail address:} george.jackeli@epfl.ch}

60: \begin{abstract}

61: 

62: I review  the microscopic spin-orbital Hamiltonian and ground state properties

63: of spin one-half spinel oxides with threefold $t_{2g}$ orbital degeneracy.

64: It is shown that for any orbital configuration  a ground state of corresponding 

65: spin only  Hamiltonian is infinitely degenerate in the classical limit.

66: The extensive classical degeneracy is lifted by the quantum nature of the spins, 

67: an effect similar to order-out-of-disorder phenomenon by quantum fluctuations.  This drives  the system to a non-magnetic 

68: spin-singlet dimer manifold with a residual degeneracy due to relative orientation of dimers. The magneto-elastic mechanism 

69: of lifting the ``orientational'' degeneracy is also briefly reviewed.

70: \end{abstract}

71: \begin{keyword}

72: Spin-Orbital models \sep spinel oxides \sep geometrical frustration

73: % keywords here, in the form: keyword \sep keyword

74: 

75: % PACS codes here, in the form: \PACS code \sep code

76: \PACS 75.10.Jm \sep 75.30.Et 

77: \end{keyword}

78: \end{frontmatter}

79: \section{Introduction}

80: The pyrochlore lattice which is composed by corner sharing tetrahedra, 

81: is known to be the most frustrated lattice existing in the nature.

82: For the Heisenberg antiferromagnet on  pyrochlore lattice the order-out-of-disorder mechanisms are inactive \cite{pyrtheo1} and

83: such a spin system would remain liquid

84: down to the lowest temperatures \cite{pyrtheo2}.

85: 

86: 

87: 

88: 

89: 

90: However, in real compounds magnetic ions, forming a frustrated lattice, often  possess an orbital degeneracy

91: in addition to the spin one.  In such cases, orbital degrees of freedom are also incorporated

92: in the superexchange theory and the systems are described by means of an effective 

93: Kugel-Khomskii type spin-orbital model \cite{KK}. 

94: The exchange interaction between magnetic moments on a bond depends now on the orientation

95: of occupied electronic orbitals with respect to that bond. The physical behavior of such systems is expected to

96:  be drastically different from that of pure spin models, as the occurrence

97: of an orbital ordering  can modulate the spin exchange and  partially or

98: fully release the geometrical degeneracy of the underlying lattice.

99: 

100: The compounds that we have in mind  are transition metal (TM) 

101: spinels.  The systems which  are the subject of intense experimental 

102: and theoretical activity \cite{spinel}. 

103: All spinels have the general formula AB$_2$O$_4$ and here we will be dealing with situation

104: when B sites with octahedral coordination are occupied by magnetic TM ions with orbital degeneracy.

105: The most interesting feature of the spinel structure  is the fact that the B ions  form a  highly frustrated 

106: pyrochlore lattice [See Fig.\ref{spinel}].

107: These systems thus give the unique possibility to explore how the natural tendency of correlated systems to develop

108: magnetic and orbital is effected by geometrical frustration.

109: 

110: The degeneracy of the $d$-shell of TM ions is not fully  lifted by the ligand field with octahedral symmetry

111: and local electronic structure is composed by high energy $e_g$ doublet and low energy $t_{2g}$ triplet.

112: In the case of partial filling of the $t_{2g}$ manifold, ($d_{xy}$, $d_{xz}$, and $d_{yz}$ orbitals), 

113: one encounters with the  threefold orbital degeneracy  in $d^1$ and $d^2$ systems.

114: The first case corresponds to Ti$^{3+}$ spinels, such as MgTi$_2$O$_4$ \cite{Tiex1,Tiex2},

115: and the second case to V$^{3+}$ spinels  AV$_2$O$_4,$ where A=Zn, Mg, Cd \cite{Vex1,Vex2}.

116: Although both cases are described by similar Hamiltonians the physics of $S=1/2$ and $S=1$ systems are drastically different.

117: In what follows, I focus on  the  spin one-half $d^1$ system and refer the readers to Refs.\cite{V1,V2,V3} for the

118: discussion of $d^2$ systems. 

119: 

120: The theoretical studies of the effective model for titanium spinels have been performed in Refs. \cite{Ti1,Ti2}.

121: Here, I briefly review the obtained results, however, within a different and complimentary scheme of reasoning.

122: 

123: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

124: \begin{figure}

125: \centerline{\epsfig{file=fig1.eps,width=5.0cm}}

126: %\vspace{-0.7cm}

127: \caption{The pyrochlore lattice formed by B sites in the spinel structure AB$_2$O$_4$: only B (colored circles) and O 

128: (white circles) sites are shown.}

129: \label{spinel}

130: \end{figure}

131: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

132: 

133: 

134: \section{Effective spin-orbital Hamiltonian}

135:  We assume that the low-temperature insulating phase of MgTi$_2$O$_4$ is of Mott-Hubbard type.

136: However, the alternative description of the insulating phase starting from the band picture 

137: is also possible and is given in Ref. \cite{KM}. 

138: 

139: In the Mott phase, an effective low energy spin-orbital Hamiltonian can be derived with the standard second-order superexchange theory

140: and has been reported in Refs.\cite{Ti1,Ti2}. The peculiarity of the spinel structure is due to the fact 

141:  that electron transfer between the nearest-neighbor (NN) sites of B-sublattice is 

142: governed by the direct  $dd\sigma$ overlap of $t_{2g}$ orbitals.

143: The $dd\sigma$ overlap  in $\alpha\beta$ plane

144: connects only  the corresponding orbitals of the same $\alpha\beta$ type.

145: The pyrochlore lattice, formed by B-ions, can also be viewed as a collection of crossing chains 

146: running in $xy$, $yz$, and $xz$ directions, as seen in Fig. 1.

147: Along the bond, for example in $xy$ direction, only the diagonal overlap between $xy$ orbitals are nonzero. 

148: Therefore the total number of electrons in each orbital state at a given site is

149: a conserved quantity and the orbital part of the effective Hamiltonian

150: $H$ has no dynamics. The orbital degrees are thus Potts-like $Z_3$

151: static variables. The spin-orbital Hamiltonian  has the following form:

152: \begin{eqnarray}

153: H&=&4J_{\rm AF}\sum_{\langle ij\rangle} {\big [}\vec S_i\cdot \vec S_{j}-\frac{1}{4} {\big ]}O^{\rm OF}_{ij}

154: \nonumber\\

155: &-&{\sum_{\langle ij\rangle}

156: \big [}J_{\rm O} +J_{\rm F} \vec S_i\cdot \vec S_{j}{\big ]} O^{\rm OAF}_{ij}

157: \label{eq1}

158: \end{eqnarray}

159: where the sum is over pairs of nearest-neighbor sites of the pyrochlore lattice.

160: The first term describes antiferromagnetic coupling between the NN spins and is active when the bond is occupied by the

161: corresponding orbitals of the same type [see Fig.2A]. For the bond $ij$ in $\alpha\beta$-plane

162: we have  ${O}^{\rm OF}_{ij}=P_{i,\alpha\beta}P_{j,\alpha\beta}$, where $P_{i,\alpha\beta}$ is a projector operator and is 

163: equal to 1 when $\alpha\beta$ orbital on site $i$ is occupied and is zero otherwise.

164: When the bond $ij$ in $\alpha\beta$-plane is occupied by two different orbitals, one of them being of

165: $\alpha\beta$ type [see Fig.2B], then there is gain of energy $J_{\rm O}$ , independent of spin configuration on the bond.

166: This is given by the second term of the Hamiltonian. However, in the same situation, the local Hund's coupling

167: $J_H$ favors the parallel orientation of the spins and introduces a weak ferromagnetic (FM) coupling $J_{F}$

168:  between them, described by the last term in the Hamiltonian. Along the bond in $\alpha\beta$-plane  

169: $O^{\rm OAF}_{ij}=P_{i,\alpha\beta}(1-P_{j,\alpha\beta})+P_{j,\alpha\beta}(1-P_{i,\alpha\beta})$. There is a local constraint

170: $P_{i,xy}+P_{i,xz}+P_{i,yz}=1$ as we have one electron at each site.

171: The coupling constants can be expressed in terms of $J=t^2/U$,  defining the  energy scale of the problem, and parameter $\eta=J_{\rm H}/U$ 

172: measuring the strength of the Hund's coupling with respect of local Coulomb repulsion $U$. For TM ion $\eta=J_{\rm H}/U\ll 1$ 

173: and is of the order of 0.1. In this limit  we have $J_{\rm AF}\simeq J_{\rm O}\sim J$, and $J_{\rm F}\sim \eta J$ \cite{Ti1,Ti2}.

174: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

175: \begin{figure}

176: \centerline{\epsfig{file=fig2.eps,width=6.0cm,angle=0}}

177: %\vspace{-0.7cm}

178: \caption{Orbital arrangements on (A) antiferromagnetic and (B) ferromagnetic bonds.}

179: \label{bonds}

180: \end{figure}

181: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

182: \section{The ground state phases}

183: We start our analysis of the ground properties from well defined realistic limit $\eta=0$.

184: In this case the Hamiltonian Eq.(\ref{eq1}) simplifies and takes the following form:

185: \begin{eqnarray}

186: {\cal H}=E_{S}+4J\sum_{\langle ij\rangle} {\big [}\vec S_i\cdot \vec S_{j}+\frac{1}{4}{\big ]}O^{\rm OF}_{ij}

187: \label{eq2}

188: \end{eqnarray}

189: where $E_S=-2JN$ is the constant energy shift due to spin uncorrelated virtual fluctuations and $N$ is the number of sites.

190: In deriving Eq.(\ref{eq2}) we have used the fact that each lattice site has nearest-neighbors in all three directions $xy$, $xz$, and $yz$, and the constraint for projection operators given above.

191: 

192: The first observation is, that in the limit $\eta=0$ the only bonds occupied with corresponding same type orbitals give the contribution

193: to the energy. The spin exchanges on such bonds are antiferromagnetic. Different orbital configurations

194: will have different number and pattern of AFM exchange of underlining spin subsystems, and, in principle, different ground state energy. 

195: However, for classical Neel type configuration of spins on such  bonds $\langle\vec S_i\cdot \vec S_{j}\rangle=-1/4$ and

196: the second term in  Eq.(\ref{eq2}) vanishes. It thus follows that for any 

197: orbital pattern the spin subsystem have the same classical energy and the classical ground state is infinitely degenerate. 

198: 

199: 

200: 

201: As we have noted above 

202: each occupied orbital has finite overlap only along corresponding chain. It is easily verified that interacting AFM bonds

203: can only be connected along the straight lines. Therefore, for each orbital configuration, 

204: the spin-subsystem can be viewed as a collection of decoupled finite (or infinite) antiferromagnetic spin one-half chains. 

205: The deviation  from classical

206: value $\langle\vec S_i\cdot \vec S_{j}\rangle=-1/4$ on AFM bond can be significant for spin one-half 

207: one dimensional objects. For antiferromagnetically coupled quantum spins the ground state expectation value 

208:  $\langle\vec S_i\cdot \vec S_{j}\rangle$ is always smaller than $-1/4$ and in this case we have finite energy gain due the second term of the Hamiltonian

209: Eq. (\ref{eq2}).

210:  The quantum energy per bond depends on the length of such spin chains. With increasing the size of AFM chain 

211: we have more bonds to gain the energy, however the amount of energy we gain decreases. 

212: The maximum energy gain per bond is for the cluster of two spins coupled into singlet state, for which we have

213:  $\langle\vec S_i\cdot \vec S_{j}\rangle=-3/4$. It then appears that the minimum energy configuration corresponds to such

214: an orbital pattern for which isolated non-interacting spin-singlet dimers are formed \cite{Ti1,jack}.

215: \begin{figure}

216: \epsfysize=136mm

217: \centerline{\epsffile{fig3.eps}}

218: \caption{Two different coverings of the unit cubic cell through dimers. 

219: Locations of singlets are represented by thick links.

220: Different numbers correspond to inequivalent tetrahedra.}

221: \label{dimercov}

222: \end{figure}

223: The quantum nature of spins removes the spin  degeneracy and drives the system to spin-singlet nonmagnetic dimer manifold.

224: However, the latter is  highly degenerate with respect of dimer orientations.

225: One of the possible dimer coverings of the lattice is shown in Fig. 3a.

226: The dimers in the limit $\eta=0$ are noninteracting and the effective dimensionality of the systems is zero.

227: In the present case, as in a spin-Peierls system,  the  increase of magnetic energy gain due to the shortening of strong bonds

228:  outweights the increase in elastic energy due to the distortion of lattice. 

229: Therefore, each type of dimer covering induces corresponding distortion of the lattice and different distortion pattern will cost

230: different elastic energy. Therefore, the ``orientational'' degeneracy of dimer phase can be lifted by the elastic energy cost.

231: We have shown in 

232: Ref.\cite{Ti1,Ti2}  that  magneto-elastic interaction indeed lifts the ``orientational'' degeneracy 

233: and stabilizes the dimer pattern leading to the minimal enlargement of the unit cell.

234:  This generates a condensate of dimers in a valence bond crystal state,

235: forming one dimensional dimerized helical chains, indicated by arrows in Fig. 3b, running around the tetragonal $c$-axis.

236: Such a dimerized pattern has been  actually observed in MgTi$_2$O$_4$ \cite{Tiex2}.

237: There is a peculiar orbital ordering in the dimerized  phase: a ferro-type 

238: order along the helices  with antiferro-type order between them (see Fig. 3b).

239: 

240: Finally, let us briefly comment on the effect of finite Hund's coupling $\eta\not =0$. 

241: The latter induces the ferromagnetic coupling, $J_{\rm F}\simeq \eta J$,

242: between the spins belonging to different dimers. For  FM coupling much smaller than the binding energy of the spins into spin-singlet state

243: $J_{FM}\ll \Delta\simeq 4J$, the dimer state is stable against weak interdimer coupling. With increasing $\eta$ one finds only one phase transition, 

244: presumably first order, from singlet-state to a ferromagnetic state with a different orbital ordering \cite{Ti1,Ti2}.

245: The dimer state and ferromagnetic spin order are only possible ground state phases of spin-orbital Hamiltonian Eq.(\ref{eq1}).

246: 

247: \begin{ack}

248: I am very grateful to S. Di Matteo, C. Lacroix, and N. Perkins for  their collaborations and numerous discussions on this topic.

249: The support  by GNSF under the Grant No.06-81-4-100 is acknowledged.

250: \end{ack}  

251:  

252: \begin{thebibliography}{9}

253: \bibitem{pyrtheo1} 

254: R. Moessner and J. T. Chalker, Phys. Rev. Lett. {\bf 80}, 2929 (1998); 

255: \bibitem{pyrtheo2} B. Canals and C. Lacroix, Phys. Rev. Lett. {\bf 80}, 2933 (1998).

256: \bibitem{KK}

257:  K. I. Kugel and D. I. Khomskii, 

258: Usp. Fiz. Nauk {\bf 136}, 621 (1982) [Sov. Phys. Usp. {\bf 231}, 25 (1982)].

259: \bibitem{spinel} For a recent review see P.G. Radaelli, New Journal of Physics {\bf 73}, 53 (2005).

260: \bibitem{Tiex1}

261: M. Isobe and Y. Ueda, J. Phys. Soc. Jpn. {\bf 71}, 1848 (2002).

262: \bibitem{Tiex2}

263: M.Schmidt {\it et al}., S. W. Cheong, Phys. Rev. Lett. {\bf 92}, 056402 (2004).

264: \bibitem{Vex1}

265: M. Reehuis {\it et al}.,  Eur. Phys. J. B {\bf 35}, 311 (2003). 

266: \bibitem{Vex2}

267: S.-H. Lee {\it et al}.,  Phys. Rev. Lett. {\bf 93}, 156407 (2004).

268: \bibitem{V1}   H. Tsunetsugu and Y. Motome, Phys. Rev. B {\bf 68}, 060405(R) (2003);

269: Y. Motome and H. Tsunetsugu, {\it ibid.} {\bf 70}, 184427 (2004).

270: \bibitem{V2} O. Tchernyshyov, Phys. Rev. Lett. 93, 157206 (2004).

271: \bibitem{V3} S. Di Matteo, G. Jackeli, and N. B. Perkins, Phys. Rev. B {\bf 72}, 020408(R) (2005). 

272: \bibitem{Ti1} S. Di Matteo, G. Jackeli, C. Lacroix, and N. B. Perkins, Phys. Rev. Lett. {\bf 93}, 077208 (2004).

273: \bibitem{Ti2} S. Di Matteo, G. Jackeli, and N. B. Perkins, Phys. Rev. B {\bf 72}, 024431 (2005).

274: \bibitem{KM}  D. I. Khomskii and T. Mizokawa, Phys. Rev. Lett. {\bf 94}, 156402 (2005). 

275: \bibitem{jack} The rigorous proof will be given elsewhere.

276: \end{thebibliography}

277: \end{document}

278: 

279: 

280: