## Abstract

The spin glass state and the phase transitions into it have recently drawn much theoretical attention (1–9). In this phase the magnetic moments are supposed to be frozen but pointing in random directions, due to a competition between e.g. quenched ferromagnetic and antiferromagnetic interactions. In many systems of experimental interest (10–13) the magnetic ions are embedded in a metallic matrix and interacting with a RKKY coupling which becomes random in magnitude and sign due to the distribution of the interion spacing. However, in many theoretical models simpler short range random exchange interactions are assumed (2–9). At the present stage the theories appear not to be conclusive due to the use of mean field theory which neglects critical fluctuations and difficulties with the n = 0 (2,14) trick at low temperatures (7,8). The more recent renormalization group calculations (6) have used an expansion around d = 6 (d is the number of space dimensions) which would not lead to accurate results at d = 3. In view of those difficulties it seems worthwhile to have an exactly soluble model for the spin glass transition which although not realistic enough to correctly represent the systems under consideration may nevertheless give one some new insights on the problem.

### Keywords

Anisotropy Ferro## Preview

Unable to display preview. Download preview PDF.

### References

- 1.C. Held and M.W. Klein, Phys. Rev. Lett. 35, 1783 (1975) contains many previous references; see also M.W. Klein, unpublished.CrossRefADSGoogle Scholar
- 2.S.F. Edwards and P.W. Anderson, J. Phys. F 5, 965 (1975).CrossRefADSGoogle Scholar
- 3.D. Sherrington and B.W. Southern, J. Phys. F 5, L49 (1975).CrossRefADSGoogle Scholar
- 4.K.H. Fischer, Phys. Rev. Lett. 34, 1438 (1975) and Solid State Comm. 18, 1515 (1976).CrossRefADSGoogle Scholar
- 5.D. Sherrington and S. Kirkpatrick, Phys. Rev. Lett. 35, 1792 (1975).CrossRefADSGoogle Scholar
- 6.A.B. Harris, T.C. Lubensky and J.H. Chen, Phys. Rev. Lett. 36, 415 (1976).CrossRefADSGoogle Scholar
- 7.For a recent review, see D. Sherrington, AIP Conf. Proc. (in press).Google Scholar
- 8a.D.J. Thouless, P.W. Anderson, E. Lieb and R.G. Palmer, unpublished.Google Scholar
- 8b.J.M. Kosterlitz, D.J. Thouless and R.C. Jones, Phys. Rev. Lett. 36, 1217 (1976).CrossRefADSGoogle Scholar
- 9.D.C. Mattis, Phys. Lett. 56A, 421 (1976).ADSGoogle Scholar
- 10.Many experiments are reviewed in J.A. Mydosh, AIP Conf. Proc. 24, 131 (1974); T.A. Fiory, AIP Conf. Proc. (in press).CrossRefADSGoogle Scholar
- 11.C.N. Guy, J. Phys. F5, L242 (1975).CrossRefADSMathSciNetGoogle Scholar
- 12.D.E. Murnick, A.T. Fiory and W.J. Kossler, Phys. Rev. Lett. 36, 100 (1976).CrossRefADSGoogle Scholar
- 13a.L.E. Wenger and P.H. Keesom, Phys. Rev. B11, 3497 (1975).ADSGoogle Scholar
- 13b.B.V.B. Sarkissian and B.R. Coles, Comm. on Physics 1, 17 (1976) and references therein.Google Scholar
- 14.V.J. Emery, Phys. Rev. B11, 239 (1975).ADSGoogle Scholar
- 15.P. Lacour-Gayet and G. Toulouse, J. Phys. (Paris) 35, 425 (1974).CrossRefGoogle Scholar
- 16.Y. Imry and S. Ma, Phys. Rev. Lett. 35, 1399 (1975). In this paper the spin redefinition is also mentioned as possibly leading to a random magnetic field.CrossRefADSGoogle Scholar
- 17.A. Aharony and Y. Imry, Solid State Comm. (in press).Google Scholar
- 18.G. Grinstein, unpublished.Google Scholar
- 19.A. Aharony, Y. Imry and S. Ma, unpublished.Google Scholar
- 20.It may be argued that <τ
_{i}^{z}> = 0, on the grounds that the field conjugate to the order parameter <τ> is strictly zero, while an infinitesimal symmetry breaking field is needed, which is conventionally sent to zero after the thermodynamic limit is taken. While this remark is correct in principle, its validity may be restricted only to infinite time averages. Although the infinite time average of \(overrightarrow {{\tau_{\text{i}}}} \) is zero, the average \(overrightarrow \tau\) should point, within any reasonable macroscopic time, in a well defined direction which we choose as the z-axis.Google Scholar - 21.M.E. Fisher and J.S. Langer, Phys. Rev. Lett. 20, 665 (1968). Note that in the Born approximation, the leading singularity in the resistivity below T
_{c}is |t|^{1–α}_{n}and not |t|^{2β}_{n}due to the fact that ρ is determined by the average \( {{\text{ < SS > }}} \) correlations without the subtraction in Eq. 9. We thank M.E. Fisher and S. Alexander for discussions on this point. See also, e.g., S. Alexander, J.S. Helman and I. Balberg, Phys. Rev. B13, 304 (1976).CrossRefADSGoogle Scholar - 22.K. Binder and D. Stauffer, Phys. Lett. 57A, 177 (1976).ADSGoogle Scholar
- 23.C. Thompson, J. Phys. A9, L25 (1976).ADSGoogle Scholar
- 24.Y. Imry, Phys. Rev. B13, 3018 (1976).ADSGoogle Scholar
- 25.See, in this connection, A. Aharony, Y. Imry and S. Ma, Phys. Rev. B13, 466 (1976).ADSGoogle Scholar
- 26.The annealed case of the present model was considered by E. Domany, Phys. Lett. 49A, 339 (1974). See also, A. Aharony and B.A. Huberman, J. Phys. C9, L465 (1976).ADSGoogle Scholar
- 27.R. Harris, M. Plischke and M.J. Zuckermann, Phys. Rev. Lett. 31, 160 (1973).CrossRefADSGoogle Scholar
- 28.Chen and Lubensky (private communication) found an Ising spin glass phase for this system, using the methods of Ref. 6.Google Scholar