Lack of monotonicity in ferromagnetic Ising model phase diagrams (Q1296627)

It is well known that the ferromagnetic Ising model on \(\mathbb{Z}^d\) with \(d\geq 2\) exhibits phase coexistence only if the external magnetic field \(h\) is 0 and the temperature \(T\) is sufficiently low. If \(\mathbb{Z}^d\) is replaced by a homogeneous tree \(\mathbb{T}_b\) (where \(b\geq 2\) is the branching number), phase coexistence may occur for some values \(h\neq 0\) if \(T\) is sufficiently low. The authors construct an inhomogeneous tree \(\mathbb{T}_{b,A}\) with only two types of vertices (essentially by adding ``leaves'' to the tree \(\mathbb{T}_b\) with \(b\) large enough) for which the following is true. If \(h\) is near 1, then there are values of \(T\) for which phase coexistence occurs, but for \(T\) either large enough or small enough there is phase uniqueness. One of the ideas on which the proof is based is that at low enough temperatures the ``leaves'' added to \(\mathbb{T}_b\) provide an extra effective external field acting on the sites of \(\mathbb{T}_b\) and thus assuring phase uniqueness. If on the other hand \(T\) is large but small compared to \(b\), then an entropy effect prevents the leaves from providing a strong effective field. The authors write with great clarity and in fact the paper provides a snap introduction to Ising models on graphs and their phase diagrams.