No label defined (Q3822979)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 4096554
Language Label Description Also known as
default for all languages
No label defined
    English
    No label defined
    scientific article; zbMATH DE number 4096554

      Statements

      0 references
      0 references
      0 references
      1986
      0 references
      equilibrium properties of disordered systems
      0 references
      Ising model
      0 references
      percolation threshold
      0 references
      bond percolation
      0 references
      disordered magnets and spin glasses
      0 references
      random field Ising model
      0 references
      Gibbs states
      0 references
      dimensional reduction technique
      0 references
      [For the entire collection see Zbl 0651.00019.] In these lecture notes the author puts the main emphasis on static aspects and equilibrium properties of disordered systems. But he also gives an introductional survey of problems and achievements in related fields in disordered systems theory. After this introduction the high-temperature and/or large magnetic field properties of disordered magnets and spin glasses are discussed. The Hamiltonian is given, for example, by \[ H = -\sum_{i,j\in\mathbb{Z}^\nu} \sigma_ i J_{ij} \sigma_ j + h \sum_{i\in\mathbb{Z}^\nu} \sigma_ i, \] where \(\sigma_ j=\pm 1\), h (the magnetic field) is real and \[ J_{ij} = \begin{cases} 1 \quad&\text{with probability }p,\\ 0 \quad&\text{with probability }1-p, \end{cases} \] is a random variable. For \(h\neq 0\) there are no phase transitions in this syste m. Let \(T_ c(p)\) denote the critical temperature of this model for \(h=0\), \(T_ c(1)\) being the critical temperature of the pure Ising model. It is shown that \[ T_ c(p) < pT_ c(1) \quad\text{and}\quad T_ c(p) \begin{cases} =0 \quad&\text{for }p < p_ c,\\ >0 \quad&\text{for }p > p_ c, \end{cases} \] where \(p_ c\) is the percolation threshold of \(\nu\)-dimensional bond percolat ion. If \(T_ c(p)<T<T_ c(1)\) one encounters the famous Griffiths singularities in the disordered phase. The author shows for these temperatures that the free ene rgy f(T,h) of the system as a function of h has an essential singularity at \(h=0\) but is differentiable at \(h=0\), i.e. \(\lim_{h\to 0}\) \((\partial f/\partial h)=0\). For \(T_ c(p)<T_ 0(p)<T<T_ c(1)\), f(T,h) is shown to be \(C^{\infty}\) in h which is a recent result. The features of the high-temperature phase described above for the dilute ferro magnet appear to be generally valid for disordered magnets and spin glasses, and part of this picture will be established for a general class of models. It is also shown that ordering in dilute ferromagnets at low temperature is stable against introducing a sufficiently small density of antiferromagnetic exchange couplings, in a variety of models. However, when the density of antiferromagnetic bonds increases, frustration tends to increase as well, and curious phenomena may happen. It is known that the magnetization decreases and it may, in fact, decrease so much that its sign is opposite to the sign of the boundary condition. Curious things happen when neither ferro- nor antiferromagnetic exchange couplings percolate, but together they do. When the resulting system is highly frustrated, it is really a (dilute) spin glass. But when frustration is suppressed, i.e. frustrated loops have very low density, the system behaves like an antiferromagnet. When such a system is put into a uniform magnetic field, its behaviour is described by the random field Ising model (RFIM). This and the phase diagram of the RFIM is discussed. In particular, it is shown that if the disorder in the magnetic field is large enough, then the equilibrium state of the RFIM is unique, and connected correlations decay at arbitrary temperatures. Uniqueness and clustering hold, of course, at high temperatures for arbitrarily weak disorder. However, for \(\nu >2\), small disorder and at low temperatures long-range order is expected. This has been rigorously established for \(T=0\) and \(\nu\geq 3.\) For spin glasses much less is known. The only safely established facts concern the disordered regime and a result concerning the positivity of the entropy at \(T=0\). If the exchange couplings \(\{J_{ij}\}\) are independent, bounded random variables with \(\overline{J_{ij}}=0\) and \(\overline{| J_{ij}|^ p}\leq const.| i-j|^{-2\alpha}\) for \(p=2,3,...\), then there is no ordering (the Edwards-Anderson order parameter vanishes) for arbitrary temperatures and magnetic fields, provided \(\alpha >1\) in \(\nu =1\), \(\alpha\) \(\geq 2\) continuous internal symmetry in \(\nu =2\), and, for arbitrary \(\nu\), one strongly expects that there is a disordered high-temperature phase, provided \(\alpha >(\nu /2)\). It is expected that, in one dimension, the Ising spin glass has a transition if \(\alpha <1\). In deterministic ferromagnets the critical values of \(\alpha\) are twice as big. In the description of all possible Gibbs states of a spin glass with given \(\{J_{ij}\}\) interesting problems arise. It is conceivable that for example for \(\nu =1\) and \(1/2<\alpha<1\), exotic Gibbs states exist which could be constructed by choosing \(\{J_{ij}\}\)-dependent boundary conditions in a class of b.c. of ``measure 0''. Finally, the author derives the connection between Ising models in imaginary ma gnetic fields and branched polymers discovered by Parisi and Sourlas and the application of this connection and the dimensional reduction technique to predict the values of the critical exponents for the three-dimensional branched polymer system. (The notes have been compiled by Bovier, Glaus and the first author.) The mean field theory for branched polymers and other rigorous results would provide a natural gateway to random surfaces and their appearances in statistical physics. But there is no room for a systematic account of these matters, except for a discussion of the intricacies and subtleties that one meets when one analyses the fluctuations of the interface in the RFIM. In particular, it is discussed, why and how the Grinstein-Ma arguments fail at low temperatures, e.g. in dimension 4, and some conjectures concerning the interface of the three-dimensional RFIM.
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references