arXiv subjects
D. A. Johnston
Publications and source records attributed to D. A. Johnston.
Phase Transition Strength through Densities of General Distributions of Zeroes
A recently developed technique for the determination of the density of partition function zeroes using data coming from finite-size systems is extended to deal with cases where the zeroes are not restricted to a curve in the complex plane and/or come in degenerate sets. The efficacy of the approach is demonstrated by application to a number of models for which these features are manifest and the zeroes are readily calculable.
Information Geometry and Phase Transitions
The introduction of a metric onto the space of parameters in models in Statistical Mechanics and beyond gives an alternative perspective on their phase structure. In such a geometrization, the scalar curvature, R, plays a central role. A non-interacting model has a flat geometry (R=0), while R diverges at the critical point of an interacting one. Here, the information geometry is studied for a number of solvable statistical-mechanical models.
Information Geometry, One, Two, Three (and Four)
Although the notion of entropy lies at the core of statistical mechanics, it is not often used in statistical mechanical models to characterize phase transitions, a role more usually played by quantities such as various order parameters, specific heats or suscept ibilities. The relative entropy induces a metric, the so-called information or Fisher-Rao m etric, on the space of parameters and the geometrical invariants of this metric carry information about the phase structure of the model. In various models the scalar curvature, ${\cal R}$, of the information metric has been found to diverge at the phase transition point and a plausible scaling relation postulated. For spin models the necessity of calculating in non-zero field has limited analytic consideration to one-dimensional, mean-field and Bethe lattice Ising models. We report on previous papers in which we extended the list somewhat in the current note by considering the {\it one}-dime nsional Potts model, the {\it two}-dimensional Ising model coupled to two-dimensional quantum gravity and the {\it three}-dimensional spherical model. We note that similar ideas have been ap plied to elucidate possible critical behaviour in families of black hole solutions in {\it four} space-time dimensions.
The Information Geometry of the Spherical Model
Motivated by previous observations that geometrizing statistical mechanics offers an interesting alternative to more standard approaches,we have recently calculated the curvature (the fundamental object in this approach) of the information geometry metric for the Ising model on an ensemble of planar random graphs. The standard critical exponents for this model are alpha=-1, beta=1/2, gamma=2 and we found that the scalar curvature, R, behaves as epsilon^(-2),where epsilon = beta_c - beta is the distance from criticality. This contrasts with the naively expected R ~ epsilon^(-3) and the apparent discrepancy was traced back to the effect of a negative alpha on the scaling of R. Oddly,the set of standard critical exponents is shared with the 3D spherical model. In this paper we calculate the scaling behaviour of R for the 3D spherical model, again finding that R ~ epsilon^(-2), coinciding with the scaling behaviour of the Ising model on planar random graphs. We also discuss briefly the scaling of R in higher dimensions, where mean-field behaviour sets in.
The Information Geometry of the Ising Model on Planar Random Graphs
It has been suggested that an information geometric view of statistical mechanics in which a metric is introduced onto the space of parameters provides an interesting alternative characterisation of the phase structure, particularly in the case where there are two such parameters -- such as the Ising model with inverse temperature $β$ and external field $h$. In various two parameter calculable models the scalar curvature ${\cal R}$ of the information metric has been found to diverge at the phase transition point $β_c$ and a plausible scaling relation postulated: ${\cal R} \sim |β- β_c|^{α- 2}$. For spin models the necessity of calculating in non-zero field has limited analytic consideration to 1D, mean-field and Bethe lattice Ising models. In this letter we use the solution in field of the Ising model on an ensemble of planar random graphs (where $α=-1, β=1/2, γ=2$) to evaluate the scaling behaviour of the scalar curvature, and find ${\cal R} \sim | β- β_c |^{-2}$. The apparent discrepancy is traced back to the effect of a negative $α$.
The Information Geometry of the One-Dimensional Potts Model
In various statistical-mechanical models the introduction of a metric onto the space of parameters (e.g. the temperature variable, $β$, and the external field variable, $h$, in the case of spin models) gives an alternative perspective on the phase structure. For the one-dimensional Ising model the scalar curvature, ${\cal R}$, of this metric can be calculated explicitly in the thermodynamic limit and is found to be ${\cal R} = 1 + \cosh (h) / \sqrt{\sinh^2 (h) + \exp (- 4 β)}$. This is positive definite and, for physical fields and temperatures, diverges only at the zero-temperature, zero-field ``critical point'' of the model. In this note we calculate ${\cal R}$ for the one-dimensional $q$-state Potts model, finding an expression of the form ${\cal R} = A(q,β,h) + B (q,β,h)/\sqrt{η(q,β,h)}$, where $η(q,β,h)$ is the Potts analogue of $\sinh^2 (h) + \exp (- 4 β)$. This is no longer positive definite, but once again it diverges only at the critical point in the space of real parameters. We remark, however, that a naive analytic continuation to complex field reveals a further divergence in the Ising and Potts curvatures at the Lee-Yang edge.
Kertesz on Fat Graphs?
The identification of phase transition points, beta_c, with the percolation thresholds of suitably defined clusters of spins has proved immensely fruitful in many areas of statistical mechanics. Some time ago Kertesz suggested that such percolation thresholds for models defined in field might also have measurable physical consequences for regions of the phase diagram below beta_c, giving rise to a ``Kertesz line'' running between beta_c and the bond percolation threshold, beta_p, in the M, beta plane. Although no thermodynamic singularities were associated with this line it could still be divined by looking for a change in the behaviour of high-field series for quantities such as the free energy or magnetisation. Adler and Stauffer did precisely this with some pre-existing series for the regular square lattice and simple cubic lattice Ising models and did, indeed, find evidence for such a change in high-field series around beta_p. Since there is a general dearth of high-field series there has been no other work along these lines. In this paper we use the solution of the Ising model in field on planar random graphs by Boulatov and Kazakov to carry out a similar exercise for the Ising model on random graphs (i.e. coupled to 2D quantum gravity). We generate a high-field series for the Ising model on $Φ^4$ random graphs and examine its behaviour for evidence of a Kertesz line.
Decorating Random Quadrangulations
On various regular lattices (simple cubic, body centred cubic..) decorating an edge with an Ising spin coupled by bonds of strength L to the original vertex spins and competing with a direct anti-ferromagnetic bond of strength alpha L can give rise to three transition temperatures for suitable alpha. The system passes through ferromagnetic, paramagnetic, anti-ferromagnetic and paramagnetic phases respectively as the temperature is increased. For the square lattice on the other hand multiple decoration is required to see this effect. We note here that a single decoration suffices for the Ising model on planar random quadrangulations (coupled to 2D quantum gravity). Other random bipartite lattices such as the Penrose tiling are more akin to the regular square lattice and require multiple decoration to have any affect.
Fat Fisher Zeroes
We show that it is possible to determine the locus of Fisher zeroes in the thermodynamic limit for the Ising model on planar (``fat'') phi4 random graphs and their dual quadrangulations by matching up the real part of the high and low temperature branches of the expression for the free energy. The form of this expression for the free energy also means that series expansion results for the zeroes may be obtained with rather less effort than might appear necessary at first sight by simply reverting the series expansion of a function g(z) which appears in the solution and taking a logarithm. Unlike regular 2D lattices where numerous unphysical critical points exist with non-standard exponents, the Ising model on planar phi4 graphs displays only the physical transition at c = exp (- 2 beta) = 1/4 and a mirror transition at c=-1/4 both with KPZ/DDK exponents (alpha = -1, beta = 1/2, gamma = 2). The relation between the phi4 locus and that of the dual quadrangulations is akin to that between the (regular) triangular and honeycomb lattices since there is no self-duality.
Thin Fisher Zeroes
Biskup et al. [Phys. Rev. Lett. 84 (2000) 4794] have recently suggested that the loci of partition function zeroes can profitably be regarded as phase boundaries in the complex temperature or field planes. We obtain the Fisher zeroes for Ising and Potts models on non-planar (``thin'') regular random graphs using this approach, and note that the locus of Fisher zeroes on a Bethe lattice is identical to the corresponding random graph. Since the number of states appears as a parameter in the Potts solution the limiting locus of chromatic zeroes is also accessible.
The Yang Lee Edge Singularity on Feynman Diagrams
We investigate the Yang-Lee edge singularity on non-planar random graphs, which we consider as the Feynman Diagrams of various d=0 field theories, in order to determine the value of the edge exponent. We consider the hard dimer model on phi3 and phi4 random graphs to test the universality of the exponent with respect to coordination number, and the Ising model in an external field to test its temperature independence. The results here for generic (``thin'') random graphs provide an interesting counterpoint to the discussion by Staudacher of these models on planar random graphs.
Balls in Boxes and Quantum Gravity
Four dimensional simplicial gravity has been studied by means of Monte Carlo simulations for some time, the main outcome of the studies being that the model undergoes a discontinuous phase transition between an elongated and a crumpled phase when one changes the curvature (Newton) coupling. In the crumpled phase there are singular vertices growing extensively with the volume of the system whereas the elongated phase resembles a branched-polymer. We have postulated that this behaviour is a manifestation of the constrained-mean-field scenario as realised in the Branched Polymer or Balls-in-Boxes model. These models share all the features of 4D simplicial gravity except that they exhibit a continuous phase transition. We note here that this defect can be remedied by a suitable choice of ensemble.
Vertex Models on Feynman Diagrams
The statistical mechanics of spin models, such as the Ising or Potts models, on generic random graphs can be formulated economically by considering the N --> 1 limit of Hermitian matrix models. In this paper we consider the N --> 1 limit in complex matrix models, which describes vertex models of different sorts living on random graphs. From the graph theoretic perspective one is using matrix model and field theory inspired methods to count various classes of directed graphs. We also make some remarks on vertex models on planar random graphs (the N --> infinity limit) where the resulting matrix models are not generally soluble using currently known methods. Nonetheless, some particular cases may be mapped onto known models and hence solved.
Why Loops Don't Matter
In a series of papers we have found identical behaviour for various spin models on thin random graphs - Feynman diagrams - and the corresponding Bethe lattices. In this note we observe that in all cases the ratios of various saddle point equations in the random graph approach are identical in form to the fixed point(s) of the recursion relations which are used to solve the models on the Bethe lattice. The loops in the random graphs thus have no influence in the thermodynamic limit for such ferromagnetic spin models. We consider the correspondence explicitly for Ising and q state Potts models and also note that multi spin interaction models on cacti admit a similar correspondence with a randomised version of the cacti graphs which contain loops.
Percolation on a Feynman Diagram
In a recent paper hep-lat/9704020 we investigated Potts models on ``thin'' random graphs -- generic Feynman diagrams, using the idea that such models may be expressed as the N --> 1 limit of a matrix model. The models displayed first order transitions for all q greater than 2, giving identical behaviour to the corresponding Bethe lattice. We use here one of the results of hep-lat/9704020 namely a general saddle point solution for a q state Potts model expressed as a function of q, to investigate some peculiar features of the percolative limit q -> 1 and compare the results with those on the Bethe lattice.
Potts Models on Feynman Diagrams
We investigate numerically and analytically Potts models on ``thin'' random graphs -- generic Feynman diagrams, using the idea that such models may be expressed as the N --> 1 limit of a matrix model. The thin random graphs in this limit are locally tree-like, in distinction to the ``fat'' random graphs that appear in the planar Feynman diagram limit, more familiar from discretized models of two dimensional gravity. The interest of the thin graphs is that they give mean field theory behaviour for spin models living on them without infinite range interactions or the boundary problems of genuine tree-like structures such as the Bethe lattice. q-state Potts models display a first order transition in the mean field for q>2, so the thin graph Potts models provide a useful test case for exploring discontinuous transitions in mean field theories in which many quantities can be calculated explicitly in the saddle point approximation.
Softening Transitions with Quenched 2D Gravity
We perform extensive Monte Carlo simulations of the 10-state Potts model on quenched two-dimensional $Φ^3$ gravity graphs to study the effect of quenched connectivity disorder on the phase transition, which is strongly first order on regular lattices. The numerical data provides strong evidence that, due to the quenched randomness, the discontinuous first-order phase transition of the pure model is softened to a continuous transition.